WO2001025274A1

WO2001025274A1 - Mitofusins, fzo homologs and functional derivatives thereof

Info

Publication number: WO2001025274A1
Application number: PCT/US2000/027871
Authority: WO
Inventors: Margaret T. Fuller; Karen G. Hales; Ansgar H. Santel
Original assignee: The Board Of Trustees Of The Leland Stanford Junior University
Priority date: 1999-10-06
Filing date: 2000-10-06
Publication date: 2001-04-12
Also published as: US6953680B2; AU8003400A; US20020168673A1

Abstract

Mitofusin genes and encoded polypeptides are provided, including the Drosophila Fzo protein and its homologs from insects, other invertebrates, yeast, and vertebrates including mouse and humans. Mitofusins are large predicted GTPases with a predicted transmembrane domain, coiled-coil regions, and a C-terminal region showing a high pl. The mitofusins are the first known protein mediator of mitochondrial fusion, and mediate developmentally regulated post-meiotic fusion of mitochondria. Missense mutations that alter conserved residues required for GTP binding in other GTPases inhibit the in vivo fusogenic activity of Fzo but do not affect its localization. Fusion proteins having amino acid sequences from mitofusin transmembrane regions localize to mitochondria. Mitofusins may be used in methods of identifying anti-insect and antifungal agents. Functional derivatives of mitofusins useful for such methods are described.

Description

MITOFUSINS, FZO HOMOLOGS AND FUNCTIONAL DERIVATIVES THEREOF

INTRODUCTION Accumulation of mutations in the mitochondrial genome has been proposed as an important contributor to aging and degenerative diseases. In several cases human mitochondrial disorders have been shown to be caused by mutations or deletions of mitochondrial DNA (for a review, see Larsson et al., 1995, Ann. Rev. Genet. 29:151-178).

There is evidence for defects in energy metabolism, excitotoxicity, and for oxidative damage in the etiology of neurodegenerative diseases, including amyotrophic lateral sclerosis, Huntington's disease, Parkinson's disease, and Alzheimer's disease. It is likely that there is a complex interplay between these mechanisms. Mitochondrial DNA is particularly susceptible to oxidative stress, and there is evidence of age-dependent damage and deterioration of respiratory enzyme activities with normal aging. This may contribute to the delayed onset and age dependence of neurodegenerative diseases. Mitochondria are dynamic organelles that undergo regulated fusion in many cell types

(for reviews, see Kawano et a/., 1995, Int. Rev. Cytol. 767:49-110, and Bereiter-Hahn et al., 1994, Microscop. Res. Technique 27:198-219). Analysis of serial sections from rodent skeletal muscle, lymphocytes, liver, spinal ganglion cells, and from the yeast Saccharomyces cerevisiae has shown that all the mitochondrial material of a cell can exist as a giant branched reticulum.

Specific protein mediators that act as biomechanical triggers and/or regulate specificity and timing of membrane fusion events have been identified in a wide variety of other cellular and subcellular contexts. The best characterized biomechanically acting fusogen is influenza virus hemagglutinin (HA) that mediates fusion of endocytosed viruses to cells. Regions resembling the HA fusion peptide exist in the ADAM family proteins, which are implicated in sperm/egg and myoblast fusion.

The identification of protein mediators of mitochondrial fusion and their possible role in maintenance of mitochondrial function and genomic integrity is of great interest for diagnosis, drug screening and potential therapeutic strategies, including targeted delivery of genes, proteins and molecules to existing mitochondria. If recombination between differently mutated mitochondrial DNA molecules allows restoration of a functional copy, the ability of mitochondria to fuse may play an important role in maintenance of mitochondrial genomes. Alternatively, fusion of mitochondria may allow complementation between two mutations in different genes in the mitochondrial genome, allowing restoration of mitochondrial function even in the absence of recombination.

SUMMARY OF THE INVENTION Mitofusin genes and proteins, various functional derivatives thereof, and their nucleotide and amino acid sequences, are provided. As used herein, the term "mitofusin" indicates the Drosophila Fzo protein or any of its homologues from insects, other invertebrates, yeast, and vertebrates including mouse and humans. Mitofusins are large predicted GTPases with a predicted trans-membrane domain, coiled-coil regions, and a C- terminal region showing a high isoelectric point (pi) and a predicted coiled-coil region. The mitofusins are the first known protein mediators of mitochondrial fusion, e.g. mediating developmental^ regulated post-meiotic fusion of mitochondria in Drosophila spermatids.

In one embodiment of the invention, the mitofusin is a Mfn2 mitofusin, e.g. comprising the amino acid sequence as set forth in any one of SEQ ID NO:4; 10; or 16. In another embodiment of the invention, the mitofusin is a mammalian mitofusin, e.g. comprising the amino acid sequence set forth in any one of SEQ ID NO:6; 8; 10; 14; or 16. The human mitofusin family has been found to comprise the nucleic acid sequences as set forth in SEQ ID NO:5; 7; 9; 11 ; and 12.

DESCRIPTION OF THE SPECIFIC EMBODIMENTS

Mitofusin genes and proteins, various functional derivatives thereof, and their nucleotide and amino acid sequences, are provided. The mitofusin Fzo is the first known protein mediator of mitochondrial fusion. As used herein, the term "mitofusin" indicates the Drosophila Fzo protein or any of its homologs from cells or organisms possessing mitochondria including, by way of non-limiting example, mature and larval stages of insects such as, for example, flies, mosquitoes, butterflies, bees and moths; other invertebrates, such as, for example, worms; protozoa; yeast and other fungi; vertebrates, including birds (such as chickens and turkeys), amphibians such as frogs and salamanders, and mammals such as mice, cows, pigs, sheep, horses and humans. The term "Fzo homolog" includes all mitofusins as well as homologs of the Fzo protein derived from cells or organisms that lack mitochondria, i.e., eubacteria and archaebacteria. Nucleic acids from different organisms, cells or tissues that encode Fzo homologs and mitofusins can be identified by those skilled in the art by a variety of methods, which may be combined.

Table of Sequences

One method involves detectably labeling a nucleic acid probe, which comprises all or a portion of one of the nucleic acids that encode Fzo homologs described herein, and hybridizing the detectably labeled probe to nucleic acids prepared from the cell, tissue or organism of interest and separated on the basis of size, charge, or other physical characteristics. A widely used method of separation involves eiectrophoresis of nucleic acids on agarose of polyacrylamide gels, followed by transfer of the electrophoresed nucleic acids to an appropriate filter such as, e.g., nitrocellulose. Under high stringency hybridization conditions, Fzo homologs having at least a 80% degree of similarity to the Fzo homologs described herein may be detected by hybridization analysis; under low stringency hybridization conditions, Fzo homologs having at least a 30% degree of similarity to the Fzo homologs described herein may be detected. A plurality of probes, derived from more than one of the Fzo homologs described herein, can be used to detect Fzo homologs having a low degree of similarity (e.g., less than about 50%) to the Fzo homologs herein disclosed. Once detected, nucleic acids encoding Fzo homologs may be further purified or amplified by methods known in the art, e.g., affinity purification or polymerase chain reaction (PCR). Another method involves database searching and/or sequence comparisons using the nucleotide or amino acid sequences provided herein. As described in detail in the Examples, Fzo homologs and mitofusins comprise particular peptide motifs that can be identified by virtue of their having at least 75% identity, with no gaps, with various consensus sequences provided herein. By "having at least 75% identity, with no gaps," it is meant that an amino acid sequence in a Fzo homolog can be aligned, residue by residue, with a consensus amino acid sequence provided herein, without the need to introduce gaps in either of the sequences and with a minimum of 75% sequence identity. By way of example, consider a consensus sequence MVAFFGRTSNGKSTVINA (SEQ ID NO: 15). An example of an amino acid sequence having at least 80% identity, with no gaps, with SEQ ID NO:21 is shown below as SEQ ID NO:22:

MVAFFGRTSNGKSTVINA ( SEQ ID NO : 21 )

1 1 1 1 1 * 1 1 1 1 1 1 1 * 1 1 1 1 MVAFFLRTSNGKS HVINA ( SEQ ID NO : 22 ) ,

wherein straight lines ("|") indicate identical residues and asterisks ("^*") indicate non-identical residues. In contrast, SEQ ID NO:23 does not have at least 75% identity, with no gaps, with the consensus sequence SEQ ID NO:21 , because only about 72% (13 out of 18 residues) of the amino acid residues of SEQ ID NO:23 are identical to the consensus sequence when the two are aligned:

MVAFFGRTSNGKSTVINA (SEQ ID NO: 21)

I I I 1**1*11*11*1111 MVAFKLRHSNPKSHVINA (SEQ ID NO:23).

Similarly, SEQ ID NO:24 does not have at least 75% identity, with no gaps, with the consensus sequence SEQ ID NO:21 , because gaps (indicated as "-") must be introduced into one of the sequences in order to achieve this degree of identity when the two sequences are aligned:

MVAFFGRTSNGKSTVINA (SEQ ID NO: 21) I I I I I I I! II II II I I MVAFF-RTSNG-STVINA (SEQ ID NO:24).

Yet another method of identifying Fzo homologs involves preparing proteins en masse from a cell, tissue or organism of interest, separating them by, e.g., eiectrophoresis and transferring them to, e.g., a nitrocellulose membrane which is then probed using an antibody raised to a herein-described Fzo homolog or functional derivative (Western analysis). Antibodies specific for the consensus sequences of the invention are preferred. The antibodies bind Fzo homologs specifically or at least preferentially, and the proteins that are thereby identified may be extracted from a gel, and portions of their amino sequences determined, according to methods known in the art. The amino acid sequences so derived are used to design probes or PCR primers that can be used to isolate corresponding nudeic acids, or to prepare synthetic polypeptides that are used to immunize animals in order to produce antibodies having a high degree of specificity for the Fzo homolog of interest and which can be used to separate the Fzo homolog protein from other components in an extract by affinity purification techniques.

The provided mitofusin and Fzo homolog genes, and functional derivatives and fragments thereof (including genomic regulatory regions), have utility in the in vitro or in vivo production of mitofusin and Fzo homolog proteins, and functional derivatives and fragments thereof; as probes for the detection of nucleic acids encoding mitofusins, Fzo homologs and related gene products; and for the modulation of gene activity.

The provided mitofusin and Fzo homolog proteins, and functional derivatives and fragments thereof, have utility in enhancing, modulating or inhibiting the activity of other mitofusins and Fzo homologs, such as those that are endogenous in a cell or organism; for enhancing, modulating or inhibiting the fusion of mitochondria, both in vivo and in vitro; as immunogens to raise specific antibodies; and in drug screening for compositions that mimic or modulate mitofusin activity or expression, including functional derivatives of mitofusin and Fzo homolog proteins.

The term "functional derivative" refers to a fragment, conjugate or mutant derived from a gene or protein of interest, or combinations thereof, wherein a "fragment" is an isolated nucleic acid or polypeptide, respectively, that is derived from the gene or protein of interest. A "mutant" is a gene or protein having a sequence in which one or more nucleotides or amino acid residues, respectively, has been altered relative to the sequence of gene or protein of interest, or a sequence wherein one or more nucleotides or amino acids have been inserted into or deleted from the sequence of gene or protein of interest. Such derivatives (a) retain one or more utilities, biochemical or biological functions of the gene or protein of interest or (b) are capable of enhancing, modulating or inhibiting one or more utilities, biochemical or biological functions of the gene or protein of interest. Fragments may be obtained directly, for example, by cleaving a gene with a restriction enzyme and isolating a specific restriction fragment derived from the gene, or indirectly, for example, by choosing a portion of the nucleotide sequence of a gene (or the amino acid sequence of a protein) and chemically synthesizing an oligonucleotide (or polypeptide) having that sequence.

For a protein, functional derivatives include but are not limited to synthetic polypeptides comprising an amino acid sequence derived from the protein of interest; mutant proteins, including dominant-negative mutants (Sheppard (1994) Am. J. Respir. Cell. Mol. Biol. 7 :1-6); fusion proteins, a type of conjugate wherein a polypeptide having an amino acid sequence derived from the protein of interest is contiguous with one or more polypeptides having amino acid sequences derived from proteins other than the protein of interest; and other conjugates, such as those wherein the protein of interest or a fragment derived therefrom is structurally linked (chemically bonded) to one or more non-proteinaceous chemical moieties. For a gene, functional derivatives include but are not limited to mutant nucleic acids; nucleic acids encoding fusion proteins; probes, including synthetic oligonucleotides such as PCR primers; antisense (reverse complement) nucleic acids, including ribozymes and synthetic oligonucleotides; molecular decoys, i.e., double-stranded nucleic acids capable of binding genetic regulatory factors by virtue of having a nucleotide sequence that is recognized by such factors; and conjugates, i.e., molecules wherein the gene of interest or a fragment derived therefrom is structurally linked (chemically bonded) to one or more chemical moieties, wherein such chemical moieties are not naturally occurring nucleic acids.

A "conjugate" is a gene, protein or fragment thereof that is chemically linked to a molecular entity that is not a part of the gene or protein of interest. As will be appreciated by those skilled in the art, conjugates may have the useful property of combining, in a single molecular entity, (a) one or more utilities, biochemical or biological functions of the gene or protein of interest with (b) the chemical, physical or biological properties of the chemical moieties structurally linked thereto.

For example, nucleic acid probes can be combined with a detectable label in order to produce a probe that can be used to directly detect nucleic acids capable of hybridizing to the gene of interest or homologs thereof. One type of functional derivative of a mitofusin- encoding nucleic acid is an oligonucleotide having at least 18 nucleotides, and comprising a nucleotide sequence contained within a nucleotide sequence selected from the group consisting of SEQ ID NO:1 ; 3; 5; 7; 9; 11 ; 12; 13; 15; and 17 or the reverse complement of such a nucleotide sequence. Such oligonucleotides have the ability to specifically hybridize to nucleic acids encoding a Fzo homolog or a mitofusin.

As another example, one type of functional derivative of a Fzo homolog protein or a mitofusin is a polypeptide having at least 12 amino acids, and comprising an amino acid sequence contained within an amino acid sequence selected from the group consisting of SEQ ID NO:2; 4; 6; 8; 10; 14; 16; 18; 19; or 20. Such polypeptides can be used to immunize animals in order to produce antibodies specific for particular epitopes with a Fzo homolog or mitofusin. As another example, a fusion protein may combine a binding site for a factor, wherein the binding site corresponds to an amino acid sequence derived from a protein of interest, with an amino acid sequence that fluoresces, in order to produce a fluorescent conjugate that can be used to measure such binding and to identify molecules that enhance, modulate or inhibit such binding. The mitofusin and Fzo homolog genes, the proteins encoded thereby, and functional derivatives of such genes and proteins may be used to enhance, mediate, modulate or inhibit mitochondrial fusion, for the purposes of introducing intact or partial mitochondrial genomes (DNA molecules) or other molecules (such as mitochondrial proteins and lipids) into mitochondria in diseased host cells or tissues, either in culture or in an intact animal or individual. The use of mitofusin proteins to drive fusion of membrane-bound entities, e.g. mitochondria, in vitro enables investigation of the mechanism of mitochondrial fusion, recombination or complementation between mitochondrial DNA molecules, maintenance of mitochondrial genomes, and complementation between mitochondria. The term "complementation" refers to a situation in which, for example, two mitochondrial genomes are present in a single mitochondrion, wherein the first of such genomes is incapable of providing a mitochondrial function of interest, but the second genome nevertheless provides an amount of such function sufficient to compensate for the deficiency of the first genome.

Mitofusins are herein identified in insects (Drosophila melanogaster), yeast (Saccharomyces cerevisiae), worms (Caenorhabditis elegans), mice (Mus musculus), rats (Rattus norvegica) and humans (Homo sapiens). Representative examples of mitofusin genetic sequences are provided in SEQ ID NO:1 and 3 (Drosophila) and SEQ ID NO:5 and 9 (human). The respective encoded proteins are provided as SEQ ID NO:2 and 4 and SEQ ID NO:6 and 10, respectively. At least two distinct forms of mitofusins have been identified in mice and humans, herein termed Mfn1 and Mfn2.

The yeast, worm, mouse and human homologs share signature conserved features with Drosophila Fzo protein. All are large GTPases with a predicted trans-membrane domain, and a predicted coiled-coil region (in that order, moving from amino to carboxy termini), near their C-termini. The region of the protein C-terminal to the predicted transmembrane domain has a high pi, characteristic of proteins found in the mitochondrial matrix, while the region of the protein N-terminal to the predicted trans-membrane domain has a lower pi, characteristic of cytoplasmic proteins. These observations, coupled with the association of mitofusin protein with mitochondria at the time of fusion, with the GTPase domain exposed on the cytoplasmic face of mitochondria. The predicted GTPase activity is required for mitochondrial fusion in vivo, but not for import of the protein into or association with mitochondria.

Nucleic Acids Encoding Mitofusins

The nucleic acid sequence encoding a mitofusin may be cDNA or genomic DNA or a fragment thereof. The term "mitofusin gene" shall be intended to mean the open reading frame encoding specific mitofusin polypeptides, introns, as well as adjacent 5' and 3' non- coding nucleotide sequences involved in the regulation of expression, up to about 1 kb beyond the coding region, but possibly further in either direction. A mitofusin gene, or its corresponding cDNA, may be introduced into an appropriate vector for extrachromosomal maintenance or for integration into the genome of the host cell or organism.

The term "cDNA" as used herein is intended to include all nucleic acids that share the arrangement of sequence elements found in native mature mRNA species, where sequence elements are exons and 3' and 5' non-coding regions. Normally mRNA species have contiguous exons, with the intervening introns, when present, removed by nuclear RNA splicing, to create a continuous open reading frame encoding a mitofusin protein.

A genomic sequence of interest comprises the nucleic acid present between the initiation codon and the stop codon, as defined in the listed sequences, including all of the introns that are normally present in a native chromosome. It may further include the 3' and 5' untranslated regions found in the mature mRNA. It may further include specific transcriptional and translational regulatory sequences, such as promoters, enhancers, etc., including about 1 kb, but possibly more, of flanking genomic DNA at either the 5' or 3' end of the transcribed region. The genomic DNA may be isolated as a fragment of 100 kbp or smaller; and substantially free of flanking chromosomal sequence. The genomic DNA flanking the coding region, either 3' or 5', or internal regulatory sequences as sometimes found in introns, contains sequences required for proper tissue and stage specific expression.

Genomic regions of interest include the non-transcribed sequences 5' or 3' to a mitofusin gene. This region of DNA contains the native promoter elements that direct expression of the linked mitofusin gene. A promoter region will usually have at least about 100 nt of sequence located 5' to a mitofusin gene and will often extend 5' to include other regulatory elements.

The sequence of this 5' region may be utilized for promoter elements, including enhancer binding sites, that provide for developmental regulation in tissues where mitofusins are expressed. The tissue specific expression is useful for determining the pattern of expression, and for providing promoters that mimic the native pattern of expression. Naturally occurring polymorphisms in the promoter region are useful for determining natural variations in expression, particularly those that may be associated with disease.

Alternatively, mutations may be introduced into the promoter region to determine the effect of altering expression in experimentally defined systems. Methods for the identification of specific DNA motifs involved in the binding of transcriptional factors are known in the art, e.g. sequence similarity to known binding motifs, gel retardation studies, etc. (see, for example, see Blackwell et al. (1995) MolMed 7:194-205; Mortlock eta/. (1996) Genome Res. 6:327-33; and Joulin and Richard-Foy (1995) Eur. J. Biochem 232:620-626.

The regulatory sequences may be used to identify cis acting sequences required for transcriptional or translational regulation of mitofusin expression, especially in different tissues or stages of development, and to identify c/s-acting sequences and trans acting factors that regulate or mediate mitofusin expression. Such transcription or translational control regions may be operably linked to a mitofusin gene in order to promote expression of wild type or altered human or Drosophila mitofusins or other proteins of interest in cultured cells, or in embryonic, fetal or adult tissues, and for gene therapy. In addition coding sequences may be used to identify cis acting sequences and possible trans-acting factors that regulate or mediate import of mitofusins into or on to mitochondria. Such mitochondrial import signals may be operably linked to a mitofusin or other gene to promote import of wild-type or altered mitofusins or other proteins of interest into or onto mitochondria in vivo or in vitro.

The nucleic acid compositions of the subject invention may encode all or a part of the subject polypeptides. Double or single stranded fragments may be obtained of the DNA sequence by chemically synthesizing oligonucleotides in accordance with conventional methods, by restriction enzyme digestion, by PCR amplification, etc. For the most part, DNA fragments will be of at least 15 nt, usually at least 18 nt, more usually at least about 50 nt. Such small DNA fragments are useful as primers for PCR, hybridization screening, etc. Larger DNA fragments, i.e. greater than 100 nt are useful for production of the encoded polypeptide. For use in amplification reactions, such as PCR, a pair of primers will be used. The exact composition of the primer sequences is not critical to the invention, but for most applications the primers will hybridize to the subject sequence under stringent conditions, as known in the art. It is preferable to choose a pair of primers that will generate an amplification product of at least about 50 nt, preferably at least about 100 nt. Algorithms for the selection of primer sequences are generally known, and are available in commercial software packages. Amplification primers hybridize to complementary strands of DNA, and will prime towards each other.

The mitofusin genes are isolated and obtained in substantial purity, generally as other than an intact chromosome. Usually, the DNA will be obtained substantially free of other nucleic acid sequences that do not include a mitofusin sequence or fragment thereof, generally being at least about 50%, usually at least about 90% pure and are typically "recombinant," i.e. flanked by one or more nucleotides with which it is not normally associated on a naturally occurring chromosome.

The DNA sequences are used in a variety of ways. They may be used as probes for identifying mitofusin related genes. Between mammalian species, e.g. human and mouse, homologs have substantial sequence similarity, i.e. at least 75% sequence identity between nucleotide sequences. Sequence similarity is calculated based on a reference sequence, which may be a subset of a larger sequence, such as a conserved motif, coding region, flanking region, etc. A reference sequence will usually be at least about 18 nt long, more usually at least about 30 nt long, and may extend to the complete sequence that is being compared. Algorithms for sequence analysis are known in the art, such as BLAST, described in Altschul et al. (1990) J. Mol. Biol. 275:403-10. The amino acid sequence similarity between invertebrates and vertebrates is sufficient to identify homologous genes. Regions of more highly conserved sequence are identified through a comparison of the provided sequences. Such conserved regions are used to design degenerate oligonucleotide primers, as known in the art. Conveniently, the primers are used in an polymerase chain reaction amplification with cDNA or genomic DNA from the target organism as a substrate. The resulting amplified nucleic acid product comprises a fragment of the mitofusin from the target organism, and can be used to isolate the complete gene by various methods known in the art, including rapid amplification of cloned ends (RACE), hybridization to cDNA libraries, etc. Nucleic acids encoding all or a portion of a mitofusin homolog can also be identified by means of Southern analysis of DNA isolated from different biological sources. Total DNA can be isolated from any organism (see Ausubel, F. et al Current Protocols in Molecular Biology, 1988). The source of nucleic acids encoding mitofusins and other Fzo homologs may be any species, e.g. primate species, particularly human; rodents, such as rats and mice, canines, felines, bovines, ovines, equines, yeast, Drosophila, Caenorhabditis, etc. The isolated DNA is restriction digested, run in agarose gels and blotted onto either nitrocellulose or nylon filters. These filters are then probed with full or partial-length coding sequences of the Fzo gene or of any of the Fzo homologs or mitofusins described herein.

Nucleic acids having sequence similarity are detected by hybridization under low stringency conditions, for example, at 50°C and 10XSSC (0.9 M saline/0.09 M sodium citrate) and remain bound when subjected to washing at 55°C in 1XSSC. At low stringency hybridization and washing conditions, nucleic acids with a degree of similarity to a nucleic acid encoding a Fzo homolog as low as 30% can be identified. Sequence identity may be determined by hybridization under stringent conditions, for example, at 50°C or higher and 0.1XSSC (9 mM saline/0.9 mM sodium citrate). By using probes, particularly labeled probes of DNA sequences, one can isolate homologous or related genes. At high stringent hybridization and washing conditions, only the genes with a similarity to the probe DNA higher than 80% will be identified. These high stringent conditions consist of an overnight hybridization at 68°C in a variety of buffers (Ausubel et al., eds., Current Protocols in Molecular Biology, 1988), followed by 2 washes of 10 min each at 68°C in 2x standard saline citrate, SSC,/0.1 % SDS, one wash of 10 min at 68°C in 1xSSC/0.1 % SDS, and one wash of 5 min at 68°C. in 0.1x SSC/0. 1 % SDS). The DNA may also be used to identify expression of the gene in a biological specimen. The manner in which one probes cells for the presence of particular nucleotide sequences, as genomic DNA or RNA, is well established in the literature and does not require elaboration here. Conveniently, a biological specimen is used as a source of mRNA. The mRNA may be amplified by RT-PCR, using reverse transcriptase to form a complementary DNA strand, followed by polymerase chain reaction amplification using primers specific for the subject DNA sequences. Alternatively, the mRNA sample is separated by gel eiectrophoresis, transferred to a suitable support, e.g. nitrocellulose, nylon, etc., and then probed with a fragment of the subject DNA as a probe. Other techniques, such as oligonucleotide ligation assays, in situ hybridizations, and hybridization to DNA probes arrayed on a solid chip may also find use. Detection of mRNA hybridizing to the subject sequence is indicative of mitofusin gene expression in the sample.

The sequence of a mitofusin gene, including flanking promoter regions and coding regions, may be mutated in various ways known in the art to generate targeted changes in promoter strength, sequence of the encoded protein, etc. The DNA sequence or product of such a mutation will be substantially similar to the sequences provided herein, i.e. will differ by at least one nucleotide or amino acid, respectively, and may differ by at least two but not more than about ten nucleotides or amino acids. The sequence changes may be substitutions, insertions or deletions. Deletions may further include larger changes, such as deletions of a domain or exon. Other modifications of interest include epitope tagging, e.g. with the FLAg system, HA, etc. For studies of subcellular localization, fusion proteins with green fluorescent proteins (GFP) may be used.

Techniques for in vitro mutagenesis of cloned genes are known. Examples of protocols for site specific mutagenesis may be found in Gustin et al., Biotechniques 14:22 (1993); Barany, Gene 37:11 1-23 (1985); Colicelli et al., Mol Gen Genet 199:537-9 (1985); and Prentki et a/., Gene 29:303-13 (1984). Methods for site specific mutagenesis can be found in Sambrook et al., Molecular Cloning: A Laboratory Manual, CSH Press 1989, pp. 15.3-15.108; Weiner et al., Gene 126:35-41 (1993); Sayers et al., Biotechniques 13:592-6 (1992); Jones and Winistorfer, Biotechniques 12:528-30 (1992); Barton etal., Nucleic Acids Res 18:7349-55 (1990); Marotti and Tomich, Gene Anal Tech 6:67-70 (1989); and Zhu, Anal Biochem 177:120-4 (1989).

Such mutated genes may be used to study structure-function relationships of mitofusins, or to alter properties of the protein that affect its function or regulation. For example, constitutively active fusogens, or a dominant negatively active protein to block fusion, may be created in this manner.

Mitofusin Polypeptides The subject gene may be employed for producing all or portions of the mitofusin protein, or functional derivatives (such as fusion proteins) of mitofusin proteins. For expression, an expression cassette may be employed. The expression cassette will provide a transcriptional and translational initiation region, which may be inducible or constitutive, where the coding region is operably linked under the transcriptional control of the transcriptional initiation region, and a transcriptional and translational termination region wherein these regions are functional in the host cell of choice. These control regions may be native to a mitofusin gene, or may be derived from exogenous sources.

The expression cassette may be integrated into the genome of a host cell, or may be part of an autogenously replicating expression construct that also comprises an origin of replication and a selectable marker such as, for example, a gene that encodes a function that confers resistance to an agent that would otherwise kill or prevent the growth of the host cell (e.g., for bacterial hosts, antibacterial agents such as ampicillin, tetracycline, kanamycin, streptomycin, chloramphenicol and neomycin; for eukaryotic hosts, cytotoxic agents such as G418, blasticidin, hygromycin and Zeocin™). The peptide may be expressed in prokaryotes or eukaryotes in accordance with conventional ways, depending upon the purpose for expression. For large scale production of the protein, a unicellular organism, such as E. coli, B. subtilis, S. cerevisiae, insect cells in combination with baculovirus vectors, or cells of a higher organism such as vertebrates, particularly mammals, e.g. COS 7 cells, may be used as the expression host cells. In many situations, it may be desirable to express the mitofusin gene in eukaryotic cells, where the mitofusin gene will benefit from native folding and post-translational modifications. Small peptides can also be synthesized in the laboratory. Peptides that are subsets of a complete mitofusin sequence, usually at least about 12 amino acids in length, more usually at least about 20 amino may be used to identify and investigate parts of the protein important for function, such as the GTPase domain, mitochondrial import signals, or the coiled-coil regions, or to raise antibodies directed against these regions.

With the availability of the protein orfragments thereof in large amounts, by employing an expression host, the protein may be isolated and purified in accordance with conventional ways. A lysate may be prepared of the expression host and the lysate purified using HPLC, exclusion chromatography, gel eiectrophoresis, affinity chromatography, or other purification technique. The purified protein will generally be at least about 80% pure, preferably at least about 90% pure, and may be up to and including 100% pure. Pure is intended to mean free of other proteins, as well as cellular debris.

The use of the expressed protein for in vitro models of mitochondrial fusion is of particular interest. The protein may be used to explore conditions and sequences that are required for association of the protein with mitochondria in cell extracts, and for setting up in vitro systems to assay mitochondrial fusion. The expressed mitofusin polypeptides are used forthe production of antibodies, where short fragments provide for antibodies specific for the particular polypeptide, and larger fragments or the entire protein allow for the production of antibodies over the surface of the polypeptide. Antibodies may be raised to the wild-type or variant forms of mitofusin. Antibodies may be raised to isolated peptides corresponding to these domains, or to the native protein, e.g. by immunization with cells expressing mitofusin, immunization with liposomes having mitofusin inserted in the membrane, etc.

Antibodies are prepared in accordance with conventional ways, where the expressed polypeptide or protein is used as an immunogen, by itself or conjugated to known immunogenic carriers, e.g. KLH, pre-S HBsAg, other viral or eukaryotic proteins, or the like. Various adjuvants may be employed, with a series of injections, as appropriate. For monoclonal antibodies, after one or more booster injections, the spleen is isolated, the lymphocytes immortalized by cell fusion, and then screened for high affinity antibody binding. The immortalized cells, i.e. hybridomas, producing the desired antibodies may then be expanded. For further description, see Monoclonal Antibodies: A Laboratory Manual, Harlow and Lane eds., Cold Spring Harbor Laboratories, Cold Spring Harbor, New York, 1988. If desired, the mRNA encoding the heavy and light chains may be isolated and mutagenized by cloning in E. coli, and the heavy and light chains mixed to further enhance the affinity of the antibody. Alternatives to in vivo immunization as a method of raising antibodies include binding to phage "display" libraries, usually in conjunction with in vitro affinity maturation. Antibodies may be "humanized" to produce therapeutic agents that may be introduced into humans without triggering an undesired immune response (Gussow ef al. (1991 ), Methods Enz. 203:99-121 ). Diagnostic Uses

Biochemical studies may be performed to determine whether a sequence polymoφhism in a mitofusin coding region or control regions is associated with disease, particularly degenerative diseases associated with mitochondrial defects, e.g. amyotrophic lateral sclerosis, Huntington's disease, Parkinson's disease, and Alzheimer's disease. Disease associated polymorphisms may include mutations that alter expression level, that affect the fusogenic activity of the protein, that alter the subcellular localization of the mitofusin, etc.

Changes in the promoter or enhancer sequence that may affect expression levels of mitofusin can be compared to expression levels of the normal allele by various methods known in the art. Methods for determining promoter or enhancer strength include quantitation of the expressed natural protein; insertion of the variant control element into a vector with a reporter gene such as β-galactosidase, luciferase, chloramphenicol acetyltransferase, etc. that provides for convenient quantitation; and the like. A number of methods are available for analyzing nucleic acids for the presence of a specific sequence. Where large amounts of DNA are available, genomic DNA is used directly. Alternatively, the region of interest is cloned into a suitable vector and grown in sufficient quantity for analysis. Cells that express a mitofusin may be used as a source of mRNA, which may be assayed directly or reverse transcribed into cDNA for analysis. The nucleic acid may be amplified by conventional techniques, such as the polymerase chain reaction (PCR), to provide sufficient amounts for analysis. The use of the polymerase chain reaction is described in Saiki, et al. (1985) Science 239:487, and a review of current techniques may be found in Sambrook, et al. Molecular Cloning: A Laboratory Manual, CSH Press 1989, pp.14.2 to 14.33. Amplification may also be used to determine whether a polymoφhism is present, by using a primer that is specific for the polymorphism. Alternatively, various methods are known in the art that utilize oligonucleotide ligation as a means of detecting polymorphisms, for examples see Riley et a/. (1990) Nucl. Acids Res. 78:2887-2890; and Delahunty et a/. (1996) Am. J. Hum. Genet. 58:1239-1246.

A detectable label may be attached to a nucleic acid probe for hybridization analyses ("Southern" and "Northern"), or linked to one or more PCR primers in an amplification reaction or otherwise included in such reactions. A "detectable label" is a substance which can be covalently attached to or firmly associated with a nucleic acid probe which confers the ability to detect the probe by a given means. A suitable detectable label will not lose the quality responsible for detectability during manipulations thereof. Suitable detectable labels include radio opaque substances (including colloidal metals); chemiluminescent molecules such as digoxigenin; fluorochromes, e.g. fluorescein isothiocyanate (FITC), rhodamine, Texas Red, phycoerythrin, allophycocyanin, 6-carboxyfluorescein (6-FAM), 2',7'-dimethoxy- 4',5'-dichloro-6-carboxyfluorescein (JOE), 6-carboxy-X-rhodamine (ROX), 6-carboxy- 2',4',7',4,7-hexachlorofluorescein (HEX), 5-carboxyfluorescein (5-FAM) or N,N,N',N'- tetramethyl-6-carboxyrhodamine (TAMRA); enzymes for which colorometric substrates are available, such as alkaline phosphatase; radioactive labels, e.g. ³²P, ^P, ^S, ³H, ¹⁴C, ¹²⁵l; etc. The label may be a two stage system, where the amplified DNA is conjugated to biotin, haptens, etc. having a high affinity binding partner, e.g. streptavidin or avidin in the case of biotin, specific antibodies in the case of haptens, etc., where the binding partner is conjugated to a detectable label. The label may be conjugated to one or both of the primers. Alternatively, the pool of nucleotides used in the amplification is labeled, so as to incorporate the label into the amplification product. The amplified or cloned fragment may be sequenced by dideoxy or other methods, and the sequence of bases compared to a wild-type mitofusin sequence. Hybridization with the variant sequence may also be used to determine its presence, by Southern blots, dot blots, etc. Single strand conformational polymoφhism (SSCP) analysis, denaturing gradient gel eiectrophoresis (DGGE), and heteroduplex analysis in gel matrices are used to detect conformational changes created by DNA sequence variation as alterations in electrophoretic mobility. The hybridization pattern of a control and variant sequence to an array of oligonucleotide probes immobilized on a solid support, as described in PCT application WO 95/11995, may also be used as a means of detecting the presence of variant sequences. Alternatively, where a polymorphism creates or destroys a recognition site for a restriction endonuclease, the fragment is digested with that endonuclease, and the products size fractionated to determine whether the fragment was digested. Fractionation is performed by gel or capillary eiectrophoresis, particularly acrylamide or agarose gels.

Hybridization with a polymorphism specific probe may also be used to determine its presence, by Southern blots, dot blots, etc. The hybridization pattern of a control and variant sequence to an array of oligonucleotide probes immobilized on a solid support, as described in U.S. Patent No. 5,445,934, or in PCT application WO 95/35505, may also be used as a means of detecting the presence of variant sequences. Screening for mutations in mitofusin may be based on the functional or antigenic characteristics of the protein. Various immunoassays designed to detect polymoφhisms in mitofusin proteins may be used in screening. Where many diverse genetic mutations lead to a particular disease phenotype, functional protein assays have proven to be effective screening tools. The activity of the encoded mitofusin protein may be determined by comparison with the wild-type protein.

Antibodies specific for a mitofusin may be used in staining or in immunoassays. Samples, as used herein, include biological fluids such as semen, blood, cerebrospinal fluid, tears, saliva, lymph, dialysis fluid and the like; organ or tissue culture derived fluids; and fluids extracted from physiological tissues. Also included in the term are derivatives and fractions of such fluids. The cells may be dissociated, in the case of solid tissues, or tissue sections may be analyzed. Alternatively a lysate of the cells may be prepared.

Diagnosis may be performed by a number of methods to determine the absence or presence or altered amounts of normal or abnormal mitofusin in patient cells. For example, detection may utilize staining of cells or histological sections, performed in accordance with conventional methods. Cells are permeabilized to stain intra-mitochondrial and/or cytoplasmic molecules. The antibodies of interest are added to the cell sample, and incubated for a period of time sufficient to allow binding to the epitope, usually at least about 10 minutes. The antibody may be labeled with radioisotopes, enzymes, fluorescers, chemiluminescers, or other labels for direct detection. Alternatively, a second stage antibody or reagent is used to amplify the signal. Such reagents are well known in the art. For example, the primary antibody may be conjugated to biotin, with horseradish peroxidase- conjugated avidin added as a second stage reagent. Final detection uses a substrate that undergoes a color change in the presence of the peroxidase Alternatively, the secondary antibody conjugated to a fluorescent compound, e.g. fluorescein, rhodamine, Texas red, etc.. The absence or presence of antibody binding may be determined by various methods, including flow cytometry of dissociated cells, microscopy, radiography, scintillation counting, etc.

Determination of the subcellular localization of mitofusin through antibody binding may be used in mutation analysis to detect mutations that cause failure to express or properly localize mitofusin proteins. The assay can be performed by immunohistochemistry or immunofluorescence, where the cell sample is stained with a mitofusin specific antibody followed by labeled secondary antibodies as described above to determine whether mitofusin is properly localized in the mitochondria. Alternatively, cell lysates may be fractionated and the level of mitofusin in the mitochondrial fraction quantitated.

Other diagnostic assays of interest are based on the functional properties of mitofusin proteins. For example, a functional assay may be based on the membrane changes mediated by mitofusin gene products. Other assays may, for example, detect conformational changes, or changes in the subcellular localization of mitofusin proteins.

Modulation of Gene Expression

The mitofusin genes, gene fragments, or the encoded protein or protein fragments are useful in gene therapy to treat degenerative and other disorders involving mitochondria, including myopathies and Alzheimer's disease. Expression vectors may be used to introduce the mitofusin gene into a cell. Such vectors generally have convenient restriction sites located near the promoter sequence to provide for the insertion of nucleic acid sequences. Transcription cassettes may be prepared comprising a transcription initiation region, the target gene or fragment thereof, and a transcriptional termination region. The transcription cassettes may be introduced into a variety of vectors, e.g. plasmid; retrovirus, e.g. lentivirus; adenovirus; and the like, where the vectors are able to transiently or stably be maintained in the cells, usually for a period of at least about one day, more usually for a period of at least about several days to several weeks. The gene or mitofusin protein may be introduced into tissues or host cells by any number of routes, including viral infection, microinjection, or fusion of vesicles or mitochondria. Jet injection may also be used for intramuscular administration, as described by Furth et al. (1992) Anal. Biochem. 205:365-368. The DNA may be coated onto gold microparticles, and delivered intradermally by a particle bombardment device, or"gene gun" as described in the literature (see, for example, Tang et al. (1992) Nature 356:152-154), where gold microprojectiles are coated with the mitofusin or DNA, then bombarded into skin cells.

Antisense molecules can be used to down-regulate expression of mitofusin in cells. The antisense reagent may be antisense oligonucleotides (ODN), particularly synthetic ODN having chemical modifications from native nucleic acids, or nucleic acid constructs that express such antisense molecules as RNA. The antisense sequence is complementary to the mRNA of the targeted gene, and inhibits expression of the targeted gene products. Antisense molecules inhibit gene expression through various mechanisms, e.g. by reducing the amount of mRNA available for translation, through activation of RNAse H, or steric hindrance. One or a combination of antisense molecules may be administered, where a combination may comprise multiple different sequences.

Antisense molecules may be produced by expression of all or a part of the target gene sequence in an appropriate vector, where the transcriptional initiation is oriented such that an antisense strand is produced as an RNA molecule. Alternatively, the antisense molecule is a synthetic oligonucleotide. Antisense oligonucleotides will generally be at least about 7, usually at least about 12, more usually at least about 20 nucleotides in length, and not more than about 500, usually not more than about 50, more usually not more than about 35 nucleotides in length, where the length is governed by efficiency of inhibition, specificity, including absence of cross-reactivity, and the like. It has been found that short oligonucleotides, of from 7 to 8 bases in length, can be strong and selective inhibitors of gene expression (see Wagner et al. (1996) Nature Biotechnology 74:840-844).

A specific region or regions of the endogenous sense strand mRNA sequence is chosen to be complemented by the antisense sequence. Selection of a specific sequence for the oligonucleotide may use an empirical method, where several candidate sequences are assayed for inhibition of expression of the target gene in an in vitro or animal model. A combination of sequences may also be used, where several regions of the mRNA sequence are selected for antisense complementation. Antisense oligonucleotides may be chemically synthesized by methods known in the art (see Wagner et al. (1993) supra, and Milligan et al., supra.) Preferred oligonucleotides are chemically modified from the native phosphodiester structure, in order to increase their intracellular stability and binding affinity. A number of such modifications have been described in the literature, which alter the chemistry of the backbone, sugars or heterocyclic bases.

Among useful changes in the backbone chemistry are phosphorothioates; phosphorodithioates, where both of the non-bridging oxygens are substituted with sulfur; phosphoroamidites; alkyl phosphotriesters and boranophosphates. Achiral phosphate derivatives include 3'-O'-5'-S-phosphorothioate, 3'-S-5'-O-phosphorothioate, 3'-CH2-5'-O- phosphonate and 3'-NH-5'-O-phosphoroamidate. Peptide nucleic acids replace the entire ribose phosphodiester backbone with a peptide linkage. Sugar modifications are also used to enhance stability and affinity. The alpha-anomer of deoxyribose may be used, where the base is inverted with respect to the natural beta-anomer. The 2'-OH of the ribose sugar may be altered to form 2'-O-methyl or2'-O-allyl sugars, which provides resistance to degradation without comprising affinity. Modification of the heterocyclic bases must maintain proper base pairing. Some useful substitutions include deoxyuridine for deoxythymidine; 5-methyl-2'- deoxycytidine and 5-bromo-2'-deoxycytidine for deoxycytidine. 5- propynyl-2'-deoxyuridine and 5-propynyl-2'-deoxycytidine have been shown to increase affinity and biological activity when substituted for deoxythymidine and deoxycytidine, respectively.

As an alternative to antisense inhibitors, catalytic nucleic acid compounds, e.g. ribozymes, antisense conjugates, etc. may be used to inhibit gene expression. Ribozymes may be synthesized in vitro and administered to the patient, or may be encoded on an expression vector, from which the ribozyme is synthesized in the targeted cell (for example, see International patent application WO 9523225, and Beigelman et al. (1995) Nucl. Acids Res 23:4434-4442). Examples of oligonucleotides with catalytic activity are described in WO 9506764. Conjugates of antisense ODN with a metal complex, e.g. terpyridylCu(ll), capable of mediating mRNA hydrolysis are described in Bashkin et a/. (1995) Appl. Biochem. Biotechnol. 54:43-56.

Cell and Animal Models

The subject nucleic acids can be used to generate genetically modified non-human animals or site specific gene modifications in cell lines, for the study of mitofusin function or regulation, or to create animal models of diseases, including mitochondrial diseases, muscle myopathies, neurodegenerative disorders, and aging. The term DtransgenicD is intended to encompass genetically modified animals having a deletion or other knock-out of mitofusin gene activity, having an exogenous mitofusin gene that is stably transmitted in the host cells where the gene may be altered in sequence to produce a modified protein, or having an exogenous mitofusin promoter operably linked to a reporter gene. Transgenic animals may be made through homologous recombination, where the mitofusin locus is altered. Alternatively, a nucleic acid construct is randomly integrated into the genome. Vectors for stable integration include plasmids, retroviruses and other animal viruses, YACs, and the like. Of interest are transgenic mammals, e.g. cows, pigs, goats, horses, etc., and particularly rodents, e.g. rats, mice, etc.

Investigation of genetic function may also utilize non-mammalian models, particularly using those organisms that are biologically and genetically well-characterized, such as C. elegans, D. melanogaster and S. cerevisiae. The subject gene sequences may be used to knock-out corresponding gene function or to complement defined genetic lesions in order to determine the physiological and biochemical pathways involved in mitofusin function. Drug screening may be performed in combination with complementation or knock-out studies, e.g. to study progression of degenerative disease, to test therapies, or for drug discovery. The modified cells or animals are useful in the study of mitofusin function and regulation. For example, a series of small deletions and/or substitutions may be made in the mitofusin gene to determine the role of different domains in GTPase activity, membrane fusion, etc. Specific constructs of interest include, but are not limited to, antisense mitofusin constructs to block mitofusin expression, expression of dominant negative mitofusin mutations, and over-expression of a mitofusin gene. One may also provide for expression of the mitofusin gene or variants thereof in cells or tissues where it is not normally expressed or at abnormal times of development. In addition, by providing expression of mitofusin protein in cells in which it is otherwise not normally produced, one can induce changes in mitochondrial behavior. DNA constructs for homologous recombination will comprise at least a portion of the mitofusin gene with the desired genetic modification, and will include regions of homology to the target locus. DNA constructs for random integration need not include regions of homology to mediate recombination. Conveniently, markers for positive and negative selection are included. Methods for generating cells having targeted gene modifications through homologous recombination are known in the art. For various techniques for transfecting mammalian cells, see Keown etal. (1990) Methods in Enzymology 785:527-537.

For embryonic stem (ES) cells, an ES cell line may be employed, or embryonic cells may be obtained freshly from a host, e.g. mouse, rat, guinea pig, etc. Such cells are grown on an appropriate fibroblast-feeder layer or grown in the presence of appropriate growth factors, such as leukemia inhibiting factor (LIF). When ES cells have been transformed, they may be used to produce transgenic animals. After transformation, the cells are plated onto a feeder layer in an appropriate medium. Cells containing the construct may be detected by employing a selective medium. After sufficient time for colonies to grow, they are picked and analyzed for the occurrence of homologous recombination or integration of the construct. Those colonies that are positive may then be used for embryo manipulation and blastocyst injection. Blastocysts are obtained from 4 to 6 week old superovulated females. The ES cells are trypsinized, and the modified cells are injected into the blastocoel of the blastocyst. After injection, the blastocysts are returned to each uterine horn of pseudopregnant females. Females are then allowed to go to term and the resulting chimeric animals screened for cells bearing the construct. By providing for a different phenotype of the blastocyst and the ES cells, chimeric progeny can be readily detected.

The chimeric animals are screened for the presence of the modified gene and males and females having the modification are mated to produce homozygous progeny. If the gene alterations cause lethality at some point in development, tissues or organs can be maintained as allogeneic or congenic grafts or transplants, or in in vitro culture. The transgenic animals may be any non-human mammal, such as laboratory animals, domestic animals, etc. The transgenic animals may be used in functional studies, drug screening, etc., e.g. to determine the effect of a candidate drug on mitochondrial fusion, to test potential therapeutics or treatment regimens, etc.

Screening Assays

By providing for the production of large amounts of mitofusin protein, one can identify ligands or substrates that bind to, modulate or mimic the action of mitofusin. Drug screening identifies agents that provide a replacement or enhancement for mitofusin function in affected cells. A wide variety of assays may be used for this puφose, including labeled in vitro protein-protein binding assays, protein-DNA binding assays, protein-mitochondria or protein mitochondria fraction, electrophoretic mobility shift assays, immunoassays for protein binding, and the like. All or a fragment of the purified protein may also be used for determination of three-dimensional crystal structure, which can be used for modeling intermolecular interactions, membrane fusion, etc. Altered mitofusin molecules, either as an isolated recombinant protein, or as a genetically modified cell or animal model may be assayed to investigate structure function parameters, including but not limited to potential mitochondrial import sequences, protein interaction domains, GTPase motifs, and dominant negative acting forms of the protein.

The term "agent" as used herein describes any molecule, e.g. protein or pharmaceutical, with the capability of altering, mimicking or interfering with the physiological function of a subject mitofusin or Fzo homolog. Generally, a plurality of assay mixtures are run in parallel with different agent concentrations to obtain a differential response to the various concentrations. Typically, one of these concentrations serves as a negative control, i.e. at zero concentration or below the level of detection. Candidate agents encompass numerous chemical classes, though typically they are organic molecules, preferably small organic compounds having a molecular weight of more than 50 and less than about 2,500 daltons. Candidate agents comprise functional groups necessary for structural interaction with proteins, particularly hydrogen bonding, and typically include at least an amine, carbonyl, hydroxyl or carboxyl group, preferably at least two of the functional chemical groups. The candidate agents often comprise cyclical carbon or heterocyclic structures and/or aromatic or polyaromatic structures substituted with one or more of the above functional groups. Candidate agents are also found among biomolecules including, but not limited to: peptides, saccharides, fatty acids, steroids, purines, pyrimidines, derivatives, structural analogs or combinations thereof.

Candidate agents are obtained from a wide variety of sources including libraries of synthetic or natural compounds. For example, numerous means are available for random and directed synthesis of a wide variety of organic compounds and biomolecules, including expression of randomized oligonucleotides and oligopeptides. Alternatively, libraries of natural compounds in the form of bacterial, fungal, plant and animal extracts are available or readily produced. Additionally, natural or synthetically produced libraries and compounds are readily modified through conventional chemical, physical and biochemical means, and may be used to produce combinatorial libraries. Known pharmacological agents may be subjected to directed or random chemical modifications, such as acylation, alkylation, esterification, amidification, etc. to produce structural analogs.

Where the screening assay is a binding assay, one or more of the molecules may be joined to a label, where the label can directly or indirectly provide a detectable signal. Various labels include radioisotopes, fluorescers, chemiluminescers, enzymes, specific binding molecules, particles, e.g. magnetic particles, and the like. Specific binding molecules include pairs, such as biotin and streptavidin, digoxin and antidigoxin etc. Forthe specific binding members, the complementary member would normally be labeled with a molecule that provides for detection, in accordance with known procedures.

It may be preferable in some instances to use a functional derivative of a Fzo homolog or mitofusin in screening assays. For example, a fusion protein in which mitofusin amino acid sequences are linked to a detectable polypeptide, for example green fluorescent protein (GFP), can be directly detected without the need to add additional reagents for detection such as, e.g., antibodies. As another example, specific binding sites (such as the 6xHis tag, which binds Nickel; streptavidin or avidin polypeptides, which bind biotin; or glutathione S transferase, GST, which binds glutathione) can be connected to mitofusin polypeptides in fusion proteins.

Naturally occurring mitofusins are typically membrane-bound, at least partially hydrophobic proteins. Because hydrophilic and water-soluble proteins are easier to incoφorate into screening assays, including high throughput screening assays, in which reactions and binding events necessary for such assays take place in the aqueous phase, water-soluble functional derivatives of mitofusins have distinct advantages over hydrophobic, membrane-bound forms of mitofusins. As described in more detail in the Examples, functional derivatives of mitofusins that lack amino acid sequences required for membrane insertion are water-soluble, or at least more hydrophilic than the wild type proteins from which they are derived. Such hydrophilic/water-soluble can nonetheless retain one or more biochemical or biological activity or function of interest, and can be used in assays designed to identify agents that interfere with such activities or functions.

A variety of other reagents may be included in the screening assay. These include reagents like salts, neutral proteins, e.g. albumin, detergents, etc that are used to facilitate optimal protein-protein binding and/or reduce non-specific or background interactions. Reagents that improve the efficiency of the assay, such as protease inhibitors, nuclease inhibitors, anti-microbial agents, etc. may be used. The mixture of components are added in any order that provides for the requisite binding. Incubations are performed at any suitable temperature, typically between 4 and 40^°C. Incubation periods are selected for optimum activity, but may also be optimized to facilitate rapid high-throughput screening.

Arthropod Fzo proteins, Fzo homologs, mitofusins, nucleic acids encoding such proteins, and functional derivatives thereof, are used in screening assays to identify anti- arthropod agents that act by, for example, inducing sterility in and/or causing the death of undesirable arthropods, preferably in a species-specific manner, by modulating or inhibiting a mitofusin function or activity. By "species-specific," it is meant that such anti-insect agents have a deleterious effect on one or more species of arthropod pests but have little or no deleterious effects, or effects that are acceptable to appropriate regulatory agencies, on desirable insect species, plants and animals. For purposes of the present invention, arthropod pests include insects and arachnids selected from the orders Coleoptera, Diptera, Hymenoptera, Lepidoptera, Mallophaga, Homoptera, Hemiptera, Orthroptera, Thysanoptera, Dermaptera, Isoptera, Anoplura, Siphonaptera, Trichoptera, and Acari, particularly Coleoptera and Lepidoptera. A list of arthropod pests associated with major crop plants are described below. Fzo homologs and mitofusins obtained from such pests, functional derivatives thereof, and screening assays incoφorating such materials, are included within the scope of the present invention.

Diptera (flies and mosquitoes) and crops affected thereby include Hylemya platura, seedcorn maggot and Mayetiola destructor, Hessian fly for barley; Ps/7/a rosae, carrot rustfly. for carrot; Delia brassicae, cabbage maggot; De//a radicum, cabbage root fly for crucifers (broccoli, cabbage, cauliflower, collards); Agromyza parvicornis, corn blotch leafminer, Hylemya platura, seedcorn maggot for maize; Contarinia sorghicola, sorghum midge, for sorghum; Hylemya platura, seedcorn maggot for soybean; Tetanops myopaeformis, sugarbeet root maggot; Neolasioptera murtfeldtiana, sunflower seed midge; Scrobipalpula absoluta, tomato leafminer; Li omyza sativae, vegetable leafminer; Liriomyza trifolli, tomato leafminer; Liviomyza sativae, vegetable leaf miner; Sitodiplosis mosellana, wheat midge; Meomyza ame cana, wheat stem maggot; Hylemya coarctata, wheat bulb fly; Mayetiola destructor, Hessian fly (wheat). Lepidoptera (butterflies and moths) and crops affected thereby include: Ostπnia nubilalis, European corn borer (barley); Plutella xylostella, diamondback moth (canola); Heliothis virescens, cotton boll worm; Helicoverpa zea, cotton bollworm; Helicoverpa armigera, cotton bollworm; Spodoptera exigua, beet armyworm; Pectinophora gossypiella, pink bollworm; Artogeia rapae, imported cabbageworm (crucifers); Pieris brassicae, cabbage butterfly; T choplusia ni, cabbage looper, Mamestra configura, bertha army worm (crucifers); Spodoptera exigua, beet armyworm; Agrotis ipsilon, black cutworm; Plutella xylostella, diamondback moth; Agrotis segetum, common cutworm; Cydia pomonella, codling moth; Platynota idaeusalis, tufted apple bud moth; Endopiza viteana, grape berry moth; Diatraea grandiosella, southwestern corn borer; Elasmopalpus lignosellus, lesser cornstalk borer; Ostrinia nubilalis, European corn borer; Ostrinia furnacali's, Asian corn borer; Sesamia nonagroides, Mediterranean corn borer; Agrotis ipsilon, black cutworm; Spodoptera frugiperda, fall armyworm; Helicoverpa zea, corn earworm; Diatraea saccharalis, sugarcane borer; Spodoptera exigua, beet armyworm; Ostrinia nubilalis, European corn borer; Spodoptera eridania, southern armywonn; Ostrinia nubilalis, European corn borer; Phthorimaea operculella, potato tuberworm; Spodoptera frugiperda, fall armyworm; Diatraea saccharalis, sugarcane borer; Chilo suppressalis, asiatic rice borer; Helicoverpa zea, corn earworm; Scirpophaga sp.; Chilo partellus, sorghum borer; Elasmopalpus lignosellus, lesser cornstalk borer; Feltia subterranea, granulate cutworm; Spodoptera frugiperda, fall armyworm; Helicoverpa zea, corn earworm; Pseudoplusia includens, soybean looper; Agrotis ipsilon, black cutworm; Anticarsia gemmatalis, velvetbean cateφillar; Plathypena scabra, green cloverworm; Spodoptera exigua, beet armyworm; Ostrinia nubilalis, European corn borer; Heliothis virescens, cotton boll worm; Helicoverpa zea, cotton bollworm; Diatraea saccharalis, sugarcane borer; Agrotis ipsilon, black cutworm; Suleima helianthana, sunflower bud moth; Homoeosoma electellum, sunflower moth; Keiferia lycopersicella, tomato pinworm; Helicoverpa zea, tomato fruitworm; Manduca quinquemaculata, tomato hornworm; Spodoptera exigua, beet armyworm; Spodoptera praefica, western yellowstriped armyworm; Spodoptera ornithogalli, yellowstriped armyworm; Spodoptera eridania, southern armyworm; Spodoptera frugiperda, fall armyworm; Agrotis ipsilon, black cutworm; Peridroma saucia, variegated cutworm; Papaipema nebris, stalk borer; Trichoplusia ni, cabbage looper; Manduca sexta, tobacco hornworm; Agrotis orthogonia, pale western cutworm; Pseudaletia unipunctata, army worm; Spodoptera frugiperda, fall armyworm; Elasmopalpus lignosellus, lesser cornstalk borer; Hemiptera (Bugs) and crops affected thereby include Blissus leucopterus leucopterus, chinch bug; Acrosternum hilare, green stink bug; Euschistus servus, brown stink bug; Lygus lineolaris, tarnished plant bug; Blissus leucopterus leucopterus, chinch bug; Blissus leucopterus leucopterus, chinch bug; Acrosternum hilare, green stink bug; Blissus leucopterus leucopterus, chinch bug; Acrosternum hilare, green stink bug; lygus bug; Acrosternum hilare, green stink bug; Euschistus servus, brown stick bug; etc.

Beetles of interest include Ceutorhychus assimils, cabbage seedpod weevil; Phyllotreta cruciferae, crucifer flea beetle; Anthonomus grandis, boll weevil; Phyllotreta cruciferae, crucifer flea beetle; Phyllotreta pusilla, western black flea beetle; Diabrotica virgifera. virgifera, western corn rootworm; Diabrotica longicornis barberi, northern corn rootworm; Diabrotica undecimpunctata howardi, southern corn rootworm; Melanotus spp., wireworms; Cyclocephala borealis, northern masked chafer (white grub); Cyclocephala immaculata, southern masked chafer (white grub); Popillia japonica, Japanese beetle; Chaetocnema pulicaria, corn flea beetle; Sphenophorus maidis, maize billbug; Anthonomus eugenii, pepper weevil; Leptinotarsa decemlineata, Colorado potato beetle; Epitrixcucu metis, potato flea beetle; Hemicrepidus memnonius, wireworms; Colaspis brunnea, grape colaspis; Lissorhoptrus oryzophilus, rice water weevil; Sitophilus oryzae, rice weevil; Oulema oryzae, rice beetle; Phyllophaga crinita, white grub; Eleodes, Conoderus, and Aeolus spp., wireworms; Oulema melanopus, cereal leaf beetle; Chaetocnema pulicaria, corn flea beetle; Sphenophorus maidis, maize billbug; Epilachna varivestis, Mexican bean beetle; Zygogramma exclamationis, sunflower beetle; Bothyrus gibbosus, carrot beetle; Melanotus spp., wireworms; Leptinotarsa decemlineata, Colorado potato beetle; Epitrix hirtipennis, tobacco flea beetle; Oulema melanopus, cereal leaf beetle; Hypera punctata, clover leaf weevil; Diabrotica undecimpunctata howardi, southern corn rootworm

Hymenoptera (sawflies, ants, wasps etc) and crops affected thereby include Solenopsis milesta, thief ant; and Cephus cinctus, wheat stem sawfly.

Homoptera (whiteflies, aphids etc.) include Schizaphis graminum, greenbug; Brevicoryne brassicae, cabbage aphid; Cavariella aegopodii, carrot aphid; Aphis gossypii, cotton aphid; Pseudatomoscelis seriatus, cotton fleahopper; Trialeurodes abutilonea, bandedwinged whitefly; Brevicoryne brassicae, cabbage aphid; Myzus persicae, green peach aphid; Dysaphis plantaginea, rosy apple aphid; Rhopalosiphum maidis, corn leaf aphid; Anuraphis maidiradicis, corn root aphid; Bemisia argentifolii, silverleaf whitefly; Bemisia tabaci, sweetpotato whitefly; Brevicoryne brassicae, cabbage aphid; Myzus persicae, green peach aphid; Enipoasca fabae, potato leafhopper; Myzus persicae, green peach aphid; Macrosiphum euphorbiae, potato aphid; Paratrioza cockerelli, potato psyllid; Nephotettix nigropictus, rice leafhopper; Nilaparvata lugens; Sogatella furcifera; Laodelphaax striatellus; Rhopalosiphum maidis, corn leaf aphid; Sipha flava, yellow sugarcane aphid; Myzus persicae, green peach aphid; Empoasca fabae, potato leafhopper; Pemphigus popullivenae, sugar beet root aphid; Saccharosydne saccharivora, West Indian canefly; Sipha flava, yellow sugarcane aphid; Myzus persicae, green peach aphid; Macrosiphum euphorbiae, potato aphid; Trileurodes vaporariorum, greenhouse whitefly; Bemisia tabaci, sweetpotato whitefly; Bemisia argentifolii, silverleaf whitefly; Russian wheat aphid; Schizaphis graminum, greenbug; Macrosiphum avenae, English grain aphid; Aphis fabae, bean aphid. Orthoptera (grasshoppers crickets, and cockroaches) include Melanoplus femurrubrum, redlegged grasshopper; Melanoplus differentialis, differential grasshopper; Melanoplus femurrubrum, redlegged grasshopper; Melanoplus sanguinipes, migratory grasshopper; Melanoplus femurrubrum, redlegged grasshopper; Melanoplus differentiahs, differential grasshopper; Melanoplus femurrubrum, redlegged grasshopper; Melanoplus differenualis, differential grasshopper; Melanoplus sanguinipes, migratory grasshopper.

Thysanontera (thrips) and crops affected thereby include Thrips tabaci, onion thrips; Frankliniella fusca, tobacco thrips; Thrips tabaci, onion thrips; Anaphothrips obscurus, grass thrips; Thrips palmi, melon thrips; Thrips palmi, melon thrips; Sericothrips vanabilis, soybean thrips; Thrips tabaci, onion thrips; Frankliniella occidentakis, western flower thrips; Frankliniella fusca, tobacco thrips; Thrips tabaci, onion thrips; Frankliniella fusca, tobacco thrips.

Acari (mites and ticks) and crops affected thereby include Petrobia latens, brown wheat mite; Panonychus citri, citrus red mite; Brevipalpus lewisi, citrus flat mite; Phyllocoptrutra oleivora, citrus rust mite; Tetranychus cinnabarinus, carmine spider mite; Tetranychus urticae, twospotted spider mite; Panonychus ulmi, European red mite; Tetranchus sp, spider mite; Tetranychus urticae, twospotted spider mite; Tetranychus cinnabarinus, carmine spider mite; Tetranychus turkestani, strawberry spider mite; Tetranychus urticae, two spotted spider mite; Tetranychus urticae, two spotted spider mite; Tetranychus urticae, two spotted spider mite; Aculops iycopersici, tomato russet mite; Steneotarsonemus pallidus, cyclamen mite; Aceria tulipae, wheat curl mite.

Anti-arthropod agents identified according to the methods of the invention are used to management and control such undesirable insects for the puφoses of crop protection and improved health and hygiene for animals including mammals such as dogs, cats, minks, horses, cows, oxen, pigs and humans. Adult insects, pupae, larvae, or cultured insect cells (including for example, the Sf9 and Sf21 cell lines derived from Spodoptera frugiperda, and the HIGH FIVE™ cell line from Trichopolusia ni; these three cell lines are available from InVitrogen, Carlsbad, California) are used to confirm the biological activity of agents that are initially screened in vitro for their ability to enhance, modulate or inhibit one or more functions or activities of Fzo proteins and/or functional derivatives thereof. Other functional derivatives are used for target validation; for example, antisense nucleic acids are used to inhibit expression of a Fzo protein or Fzo homolog, and the effects of such down-regulation are measured in order to confirm that reduced activity of the Fzo protein or homolog will result in the predicted desired result, e.g., sterility, death, etc.

Fungal Fzo proteins, Fzo homologs, mitofusins, nucleic acids encoding such proteins, and functional derivatives thereof, are further used in screening assays to identify antifungal agents that act by, for example, blocking a step in the cell cycle of, promoting the loss of mitochondria from cells, and/or causing the death of undesirable fungi, preferably in a species-specific manner, by modulating or inhibiting a mitofusin function or activity. Functional derivatives of Fzo homologs and mitofusins are used for target validation; for example, antisense nucleic acids are used to inhibit expression of a Fzo homolog or mitofusin, and the effects of such down-regulation are measured in order to confirm that reduced activity of the Fzo homolog or mitofusin will result in the predicted desired result, e.g., cell cycle arrest, cytotoxicity, etc.

One group of undesirable fungi, for which there is a need for antifungal agents forthe management thereof, include fungal phytopathogens of plants. A list of fungal pests associated with major crop plants is provided below. Fzo homologs, mitofusins, and functional derivatives thereof, and screening assays incoφorating such materials from such fungi are included within the scope of the present invention.

Fungal diseases of plants include Brown stripe downy mildew; Sclerophthora rayssiae; Graminicola downy mildew, Sclerospora graminicola; Java downy mildew, Peronosclerospora maydis; Phillipine downy mildew, Peronoscler. Philippinensis; Rajasthan downy mildew, Peronosclerospora heteropogoni ; Sorghum downy mildew, Peronosclerospora sorghi; Spontaneum downy mildew, Peronosclerospora spontanea; Sugarcane downy mildew, Peronosclerospora sacchari; Rusts, Puccinia graminis f.sp. tritici; Puccinia recondita f.sp. tritici; Puccinia striiformis; Smuts, Tilletia tritici; Tilletia controversa; Tilletia indica; Ustilago tritici; Urocystis tritici; Root rots, Foot rots and Blights, Gaeumannomyces graminis; Pythium spp.; Fusarium culmorum; Fusarium graminaerum; Fusarium avenaceum; Drechslere tritici- repentis; Rhizoctonia spp.; Colletotrichum graminicola; Helminthosporium spp.; Microdochium nivale; Pseudocercosporella herpotrichoides; Mildews, Erysiphe graminis f.sp. tritici; Sclerophthora macrospora; Ear Molds, Gibberella ear mold, Gibberella zeae; G. saubinetti; Aspergillus ear rot, Aspergillus flavus; A. parasiticus; Diplodia ear rot, Diplodia maydis; D. macrospora; Fusarium ear rot, Fusarium moniliforme; F. monilif. var. subglutinans; Leaf Diseases: Eye spot, Kabatiella zeae; Gray leaf spot , Cercospora zeae-maydis; C. sorghi var. maydis; Anthracnose leaf blight, Colletotrichum graminicola; Northern corn leaf blight, Exserohilum turcicum; Southern corn leaf blight, Bipolaris maydis; Common rust, Puccinia sorghi; P. maydis; Southern rust, Puccinia polysora; Dicaeoma polysorum; Head smut, Sphacelotheca reiliana; Common smut, Ustilago maydis; Carbonum leaf spot, Helminthosporium carbonum; Stalk Rots: Fusarium stalk rot, Fusarium moniliforme; Gibberella stalk rot, G. zeae; Pythium stalk rot, Pythium aphanidermata; Anthracnose stalk rot, Colletotrichum graminicola; Glomerella graminicola; C. tucumanensis; Diplodia stalk rot, Diplodia maydis; D. zeae-maydis; Sphaeria maydis; S. zeae; Stenocarpella maydis; Macrodiplodia zeae; D. macrospora; G. saubinetti; Stewart's wilt & leaf blight, Erwinia stewartii; Miscellaneous Fungal Diseases: Septoria tritici; Septoria nodorum; etc. Mammalian Fzo proteins, Fzo homologs, mitofusins, nucleic acids encoding such proteins, and functional derivatives thereof, are further used in screening assays to identify therapeutic agents that act by, for example, preferentially promoting the loss of mitochondria from, blocking a step in the cell cycle of, or causing the death of undesirable cells such as cancer cells and viral-infected cells by modulating or inhibiting a mitofusin function or activity.

In other embodiments, mammalian Fzo homologs, mitofusins and functional derivatives thereof are used in screening assays to identify therapeutic agents useful for treating a mitochondrial disease, wherein such agents enhance a mitofusin function or activity and thereby promote intermitochondrial complementation and/or recombination. That is, it is expected that some mitochondrial deficiencies can be corrected or at least ameriolated by promoting mitochondrial fusion by agents that enhance mitofusin functions or activities. itochondrial diseases include, by way of example and not limitation, chronic neurodegenerative disorders such as Alzheimer's disease (AD) and Parkinson's disease (PD); auto-immune diseases; diabetes mellitus, including Type I and Type II; mitochondria associated diseases, including but not limited to congenital muscular dystrophy with mitochondrial structural abnormalities, fatal infantile myopathy with severe mtDNA depletion and benign "later-onset" myopathy with moderate reduction in mtDNA, MELAS (mitochondrial encephalopathy, lactic acidosis, and stroke) and MIDD (mitochondrial diabetes and deafness); MERFF (myoclonic epilepsy ragged red fiber syndrome); NARP (Neuropathy; Ataxia; Retinitis Pigmentosa); MNGIE (Myopathy and external ophthalmoplegia; Neuropathy; Gastro-lntestinal; Encephalopathy), LHON (Leber"s; Hereditary; Optic; Neuropathy), Keams- Sayre disease; Pearson's Syndrome; PEO (Progressive External Ophthalmoplegia); Wolfram syndrome DIDMOAD (Diabetes Insipidus, Diabetes Mellitus, Optic Atrophy, Deafness); Leigh's Syndrome; dystonia; stroke; schizophrenia; progressive joint disorders, such as osteoarthritis; and hypeφroliferative disorders, such as cancer, tumors and psoriasis. unctional derivatives of Fzo homologs and mitofusins are used for target validation of mitochondrial diseases; for example, antisense nucleic acids are used to inhibit expression of a Fzo homolog or mitofusin in a cultured cell line, and the biochemical or biological effects of such down-regulation are measured in order to confirm that reduced activity of the Fzo homolog or mitofusin will result in the predicted desired result, e.g., enhanced or decreased mitochondrial stability or cytotoxicity. EXAMPLES The following examples are put forth so as to provide those of ordinary skill in the art with a complete disclosure and description of how to make and use the subject invention, and are not intended to limit the scope of what is regarded as the invention. Efforts have been made to ensure accuracy with respect to the numbers used (e.g. amounts, temperature, concentrations, etc.) but some experimental errors and deviations should be allowed for. Unless otherwise indicated, parts are parts by weight, molecular weight is average molecular weight, temperature is in degrees centigrade; and pressure is at or near atmospheric.

Example 1:

Isolation and Genetic Mapping of fzo Mutations Flies were grown on standard cornmeal molasses agar media at 25^°C. All visible markers and balancer chromosomes were as described in Flybase (1996) Nucl. Acids Res.

24:53-56 (flybase.bio.indiana.edu) unless otherwise noted. The fzo¹ allele was isolated in a screen for ethyl methanesulfonate (EMS) induced recessive male sterile mutations. The fzo² allele was isolated in an EMS screen essentially as described in Lin et al. (1996) Devel.

122: 1331 -1341 , except that mutagenized chromosomes were tested for failure to complement fzo¹. One new allele, fzo², was identified from 1799 mutagenized third chromosomes screened. Deletion Mapping. The chromosomal location of the fro¹ allele was initially identified by deletion mapping. The deletions Df(3R)M95A (94D; 95A3) and Df(3R)EB6 (94C2-5;

94E3; Mohler et al. (1995) Devel. 121, 237-247). The Df(3R)P20 12 kb deletion was generated by the mobilization and imprecise excision of a lethal ry⁺ marked P element in

ΛOD7Λ 03781

94E3-5 associated with the cnc allele as follows. Flies transheterozygous for cnc and a TM3, Sb chromosome carrying the Δ2-3 transposase insert (Reuter et al. (1993) Dros.

03781*

Inf. Serv. 72, 78) were crossed to ry e bar3 tx/TM3, Sb flies. F1 cnc /ry e bar3 tx males which had lost the P(ty+) and therefore had ry eyes were crossed to e fzo¹/ TM3, Ser

03781* 1 females. F2 cnc /e fzo males were tested for the fzo phenotype, and stocks of new

03781 * alleles were recovered by crossing together male and virgin female cnc /TM3, Ser F2 siblings. From among approximately 200 ry chromosomes tested, two deletions failing to complement fzo¹ were identified. Molecular breakpoints of the smaller deletion, Df(3R)P20 were mapped by Southern hybridization analysis using standard methods. Df(3R)P20 failed to complement mutations in cnc and was therefore not homozygous viable.

Recombination Mapping. The fzo¹ mutation was mapped by recombination to 0.5 cM distal of hh^bar3 and 0.5 cM proximal of pnt⁰⁷⁸²⁵ as follows. For initial meiotic mapping, females heterozygous for an e fzo¹ chromosome and a ru h th st p" cu sr ca chromosome were crossed to ru h th st pf cu sr Bsb/TM3 or ru h th st cu e Pri ca TM3 males. Male progeny with markers indicating single recombination events were selected for each interval and individually crossed to three e fzo¹/TM3 females to score fzo and make stocks of the recombinant chromosomes with TM3. The fzo¹ mutation mapped between e and ca, roughly 9 cM distal to e.

To localize fzo¹ further, r e fzo¹/hh^bar3 tx females were crossed to e hh^bar3 tx/TM3 males and recombinant male progeny crossed individually to e fro 7TM3females to score fzo and make stocks with TM3. fzo¹ mapped 0.58±0.17 cM distal to hh¹ "³, based on 227 recombinants between hh ^ar and tx. The fzo¹ mutation was further localized by recombination with respect to restriction fragment length polymoφhisms (RFLPs) in the interval between hh^bar3 and a ry⁺ marked P element insertion in the pointed gene associated with the pnt⁰⁷⁸²⁵ allele. Briefly, females transheterozygous for ry⁵⁰⁶ hh^bβr3 fzo¹ (parental chromosome 1 ) and ry⁵⁰⁶ pnt⁰⁷⁸²⁵ (parental chromosome 2) were crossed to ry⁵⁰⁶e hh^bar3 tx/TM3 males. Recombinant F1 Sb⁺ progeny that were phenotypically either hh^bar3~ ry⁺ or hh^bar3+ ry appeared at a frequency of 31 per 2995 flies, indicating that hh^bar3 and pnt⁰⁷⁸²⁵ were 1.04±0.18 cM apart. Males representing each of these two reciprocal recombinant classes were crossed individually to e fzo¹/TM3 females to score fzo on the recombinant chromosomes and to construct stocks with TM3. The fzo¹ mutation mapped in the center of the interval, 0.5 cM distal to hh^bar3 and 0.5 cM proximal to pnt⁰⁷⁸²⁵, in agreement with complementation of fzo¹ by Df(3R)EB6 but not by Df(3R)M95A.

RFLP Mapping. The fzo¹ mutation was further localized by RFLP mapping to an 11 kb region defined by polymoφhic Xbal and EcoRI sites on an existing genomic phage walk as follows. RFLPs between the two parental chromosomes were identified using Southern blot analysis. Standard molecular biology techniques were used throughout this work, and DNA fragments were purified with a QIAquick kit (QIAGEN) as per the manufacturer's instructions. Genomic DNA was isolated from flies homozygous for parental chromosome 1 as well as from flies transheterozygous for parental chromosomes 1 and 2 (the P element insertion on parental chromosome 2 was a recessive lethal). The parental genomic DNAs were digested separately with 34 standard four-, five-, and six-cutter restriction enzymes, electrophoresed, blotted, and hybridized with several different radiolabeled DNA fragments from the distal third of a 90 kb genomic walk (Mohler et al. (1991) Mech. Devel. 34:3-10) . At least one RFLP between the two parental chromosomes was detected for each probe. Genomic DNA was isolated from 35 recombinant stocks (either from homozygotes or from flies transheterozygous for the recombinant chromosome and parental chromosome 1). These DNAs were digested with the restriction enzymes previously shown to have polymoφhic sites in the region, blotted, and probed with the appropriate fragments. Analysis of linkage between these molecular markers and fzo¹ defined the region of the fzo¹ mutation to be an 11 kb region between a polymoφhic Xba\ site at +11 and a polymoφhic EcoRI site at +22 on the genomic phage walk of Mohler et al. (1991) . The Xba\ and EcoRI RFLPs segregated from fzo¹ in one and two of the 35 recombinants, respectively.

Example 2:

Fzo is Required for Spermatogenesis and Male Fertility in Drosophila

Fertility was determined by placing individual flies in yeasted vials with three males or virgin females as appropriate and scoring the presence or absence of larvae after seven days at 25°C. Progeny of tested females were counted. At least ten individual males and four individual females for each of the allelic combinations fzo¹ /fzo¹ , fzo¹ /fzo², fzo¹/Df(3R)P20, and fzo²/Df(3R)P20, as well as heterozygous sibling classes fzo TM3 or Df(3R)P20/TM3 were tested. (The fzo² chromosome carried a secondary lethal and could not be made homozygous.)

The effects of fzo¹ and fro² mutations on viability were tested by crossing five individual females of each of the above allelic combinations to Df(3R)P20/ TM3 males at 25^°C and comparing numbers of fzo/Df progeny to total progeny. Progeny were collected for nine days after eclosion began. Two EMS-induced alleles, fzo¹ and fzo², were characterized as described above. fzo¹ /fzo¹, fzo¹ /fzo², fzo¹/Df(3R)P20, and fzo²/Df(3R)P20 flies showed identical phenotypes, suggesting that both mutations are strong loss of function alleles. The severity of the phenotype was consistent among all spermatids in all testes observed. The fzo¹ and fzo² mutations did not noticeably affect female fertility or overall viability of the animal.

Mitochondria undergo dramatic morphogenetic changes during spermatid differentation in Drosophila (Fuller, 1993, "Spermatogenesis" In The Development of Drosophila, Bate etal., eds., Cold Spring Harbor Press, Cold Spring Harbor, NY, pp. 71-147.

In early postmeiotic spermatids, mitochondria aggregate beside each haploid nucleus and fuse into exactly two giant mitochondrial derivatives that wrap around each other to form a spherical structure called a Nebenkern. The Nebenkern resembles an onion slice when viewed in cross section by transmission electron microscopy (TEM); the term "onion stage" thus refers to early round spermatids.

Spermatogenesis in fro mutant male flies was examined by light and electron microscopy as follows. For light microscopy, testes from various fzo genotypes as well as from wild type Oregon R flies were dissected in TB1 buffer (7 mM K₂HPO₄, 7 mM KH₂PO₄ [pH 6.7], 80 mM KCI, 16mM NaCI, 5mM MgCI₂, 1% PEG-6000), opened with forceps to allow spillage of contents, placed under cover slips, and examined immediately by phase contrast microscopy with a Zeiss Axiophot microscope. To check for a mitochondrial membrane potential, 10 μg/ml rhodamine 123 was included in the dissection buffer and samples examined under epifluorescence.

For ultrastructural analysis by transmission electron microscopy, testes were dissected in TB1 and immediately placed in fixative (2% glutaraldehyde, 1 % parafoπnaldehyde, 0.1 M sodium phosphate or sodium cacodylate buffer [pH 7]). After overnight fixation, samples were washed in 0.1 % phosphate or cacodylate buffer for 15 minutes and stained with 1 % osmium tetroxidein the same buffer for two hours. Testes were washed three times in water, stained for 1 hour in 1 % uranyl acetate, washed three times in water and dehydrated through an ethanol series (30%, 50%, 70%, 95%, 100%). After five minutes in 1 :1 ethanol: propylene oxide and five minutes in propylene oxide, samples were embedded in Spurr's resin and polymerized overnight at 60^°C. Thin sections (80-90 nm) were cut with a Reichert-Jung microtome, placed on Formvar-coated slot grids, and examined on a Phillips 410 transmission electron microscope. Mutations in fro result in defects in mitochondrial fusion during Nebenkern formation in post-meiotic early round spermatids. In wild type testes, mitochondria in each haploid spermatid aggregate beside the nucleus and fuse into a Nebenkern, a phase-dark spherical structure consisting of two interwrapped mitochondrial derivatives as viewed by TEM. In fro mutant males, mitochondria aggregate in early haploid spermatids, forming somewhat misshapen Nebenkems as viewed by phase contrast microscopy, but fail to fuse into two giant mitochondria. Instead, many smaller mitochondria appear to wrap around each other at the onion stage as viewed by TEM. Despite the prior defects in fusion, mitochondria unfurl and elongate in fro mutants. At the early elongation stage, unfurling mitochondria in fzo mutants appear fragmented compared to wild type as seen by phase contrast microscopy. At the late elongation stage, two elongating mitochondrial derivatives per spermatid are seen in wild type. In contrast, many elongating mitochondria are associated with each haploid nucleus in fzo. Cross sections through wild type elongating spermatids viewed by TEM reveal two mitochondrial derivatives associated with each axoneme; the major derivative contains electron dense paracrystalline material while the minor does not. Each axoneme in fro mutant males is associated with a large number of mitochondrial derivatives, with roughly half containing the paracrystalline material characteristic of major derivatives.

The defects in fro mutant spermatids thus appear to be specific for mitochondrial fusion, as mitochondrial aggregation, membrane wrapping, and elongation all occur (although the unfused mitochondria do not elongate the full axonemal length because they lack sufficient membrane material). In addition, spermatid mitochondria in fro mutants take up the dye rhodamine 123 in amounts similar to wild type, indicating presence of a membrane potential and suggesting that fro mutations probably do not grossly affect respiration.

Example 3:

Seguence Analysis of fro cDNA Clones

Isolation of cDNAs. cDNA clones corresponding to transcripts from the fro region were isolated from a testis cDNA library as follows. Radiolabeled EcoRI restriction fragments corresponding to +5 to +22 on the genomic walk were used to probe a λ-ZAP (Stratagene) testis cDNA library. Eighteen positive cDNA clones of various lengths were shown by

Southern hybridization analysis to be from the same transcription unit. The seven longest cDNAs were 2.4 kb and had identical restriction maps. One of these was mapped to genomic DNA by Southern hybridization and, consistent with restriction mapping analysis, seemed to have no detectable introns. This cDNA was sequenced on both strands by the dideoxy chain termination method using the ABI PRISM™ dye terminator cycle sequencing system (Perkin-

Elmer). T3 or T7 primers were used on the intact cDNA or dropout subclones, and 19- or 20- mer oligonucleotides were synthesized to prime sequencing runs through any remaining gaps. Representative members of each shorter cDNA class were restriction mapped, partially sequenced, and shown to be truncated versions of the longer class.

Amino Acid Sequence Analysis. The representative fro cDNA (2399 bp) was sequenced fully on both strands and shown to have a complete open reading frame (ORF) comprising 718 amino acids, flanking AT-rich sequences and a single consensus translational start site. The cDNA ORF nucleotide sequence was translated in silico (that is, an algorithm was used to transform the nucleotide sequence into the corresponding amino acid sequence using the universal genetic code) using DNA Strider software. Analysis of the Fzo predicted protein with the BLOCKS program indicated similarity to the dynamin family in the vicinity of the P-loop, a motif found in nucleotide-binding proteins and designated as the G1 GTPase motif (Bourne et al., (1991 ) Nature 349: 117-127).

Example 4: Expression Constructs Comprising fro

To confirm that the transcription unit from which the fro cDNA was prepared corresponds to fzo, a 4 kb genomic fragment containing the candidate locus, plus 1 kb of sequence upstream and approximately 500 bp of sequence downstream of the cDNA, was introduced into flies by P element mediated germline transformation as follows. A 3 kb genomic fragment was isolated from an Xho\IXba\ digestion of phage D14 DNA

(which corresponds to +3 to +16 on the genomic walk) and subcloned into the w⁺-marked P element transformation vector pCaSpeR4 (Thummel and Pirrotta (1992) Dros. Inf. Serv. 71, 150) to make plasmid pKH2. The 2.4 kb EcoRI restriction fragment from phage D14 had previously been subcloned into pUC9 to make construct D5-19; a 1 kb Xbal/EcoRI fragment was isolated from a digestion of D5-19 and subcloned into pKH2 to make pKH3fzo⁺. The resulting 4 kb insert in the pKH3fzo⁺ rescue construct contained the genomic region of the full length fro cDNA plus approximately 1 kb 5' and 500 base pairs 3'.

Plasmid pKH3fzo⁺ and a plasmid encoding the A 2-3 constitutively active transposase (the enzyme that directs P element mobilization) were mixed in a 3:1 molar ratio to a final DNA concentration of 0.4 mg/ml and injected into y w⁶⁷ or w¹¹¹⁸ embryos. Three independent fro⁺ insertions, one on the second chromosome and two on the third, were isolated from the progeny of 32 fertile injectees. An additional insertion on the second chromosome was obtained by mobilizing one of the third chromosome insertions with the TM3, Sb Δ 2-3 transposase source.

To test for rescue, all four independent fro⁺ insertions were crossed into w; fzo mutant backgrounds by independent assortment for the second chromosome insertions and by recombination onto a hh^bar3 fzo¹ chromosome and then independent assortment for the third chromosome insertions, which were tested for rescue only of allelic combinations with fro⁷. A single copy of any of four independent transgene insertions fully restored fertility and normal mitochondrial morphogenesis to fro ¹/fzo¹ , fzo¹/Df(3R)P20, and fro Vfro² males (and also fzo²/Df(3R)P20 males, for the second chromosome insertions).

Example 5: Site-Directed Mutagenesis of fro and Retransformation with Mutant fro Transqenes To introduce mutations into the Fzo predicted GTP-binding domain, a 1.8 kb Xba\/Bam \ restriction fragment (representing the first two thirds of the fro coding region) from the pKH3fzo⁺ germline transformation construct (see above) was subcloned into BluescriptSK-. Using standard methods, mutagenesis was performed with oligonucleotides having the sequences

5'-ACCTCAAATGGAACTAGTGCCGTGATC-3'; (SEQ ID NO:25), and 5'-TACTCAACAATCTATGGGATAAG-3'; (SEQ ID NO:26). The former (SEQ ID NO:25) exchanges AA for CT at nucleotides corresponding to those at positions 369-370 in the fro cDNA, introducing a Spel restriction site and changing the encoded amino acid at position 99 from a lysine to a threonine. The latter (SEQ ID NO:26) replaces a G with a T at nucleotide 819, eliminating a C/al restriction site and changing the encoded amino acid at position 249 from an arginine to a leucine. Mutagenized constructs were selected by virtue of the above mentioned altered restriction sites. In like fashion, the mutation K110T was introduced into a fro-encoding nucleic acid.

For each mutagenized construct, a 1.8 kb Xba\lBam \ restriction fragment was subcloned back into the 10 kb Xba\/Bam \ vector fragment from plasmid pKH3fzo⁺ to create

K99T R249L two new germline transformation constructs, pKH3fzo and pKH3fzo , which were injected into fly embryos as described in preceding Examples. In both of these constructs, the Xbal and SamHI restriction sites used for the final subcloning were regenerated during the ligation step. In addition, the reading frame remained unchanged at the mutagenesis and subcloning sites, as shown by detection of mutant proteins with anti-Fzo^605"718 antibodies, which only recognize epitopes that are encoded by regions 3' to the sites of ligation and mutagenesis and which thus require translation in the correct reading frame upstream (5') therefrom. To test for rescue of the mutant phenotype by the fzo^K99T and fzo^R249L mutant transgenes, respectively, appropriate crosses were made to introduce separately the eight and five independent second chromosome transgene insertions into fzo mutant backgrounds. To test for any dominant effect, appropriate crosses were made to obtain males with one wild

K99T R24QL type copy of fro and as many as four (fro ) or six (fzo ) different copies of the mutated transgenes.

Conserved residues in the GTPase domain are required for Fzo function but not for targeting of the protein to mitochondria. Mediation of mitochondrial fusion by Fzo appears to require the predicted GTPase activity. Missense mutations (K99T and R249L) in the fro transgene that alter conserved residues (lysine-99 and arginine-249, respectively) required in other GTPases for guanine nucleotide binding were introduced into the fro transgene

K99T R249L construct. For the fro and fro transgenes, respectively, none of the eight and five independent insertions on the second chromosome (in one copy or two, for those that were viable as homozygotes) were able to restore fertility or sperm motility to fzo¹ /fzo¹ , fzo¹/Df(3R)P20, or fzo²/Df(3R)P20 mutant males. In two copies, the fzo^K99T transgene did

R249L not detectably improve the subcellular mutant phenotype, while the fro transgene appeared to allow some mitochondrial fusion. In a wildtype background, neither of the mutant transgenes impaired fertility or mitochondrial fusion; thus, neither mutant Fzo protein

K99T R249 seems to act in a negative dominant mutant in this system. Fzo and Fzo mutant proteins were properly localized to spermatid mitochondria with wild type timing of appearance and disappearance.

K99T R249L

The fro and fro mutations are recessive. Analogous mutations in mammalian dynamins cause dominant negative phenotypes when expressed in tissue culture cells; formation of macromolecular dynamin ring-shaped complexes appears to require GTP binding by all subunits. In contrast to dynamins, Fzo molecules may act individually orform complexes in which only some subunits must bind GTP for proper assembly or function. Example 6: Expression of fro in Primary Spermatocvtes RNA in situ Hybridization. For in situ hybridization to testes, digoxygenin-labeled RNA probes representing both the fzo cDNA sense and antisense strands (primed with T7 or T3 primers on a linearized cDNA construct) were made and submitted to alkaline hydrolysis (1 hour) as described in the Genius 4 RNA Labeling Kit user's guide (Boehringer Mannheim). Paraformaldehyde fixation of freshly dissected testes was as in Tautz (1989) Chromosoma 98:81-85 except no heptane or methanol was used. Subsequent treatment was as in Gonzalez (1994) "Techniques for studying mitosis in Drosophila" In The Cell Cycle: A Practical Approach, R. Brookes and P. Fantes, eds.: IRL, Oxford University Press), pp. 143-175, except 50 μg/ml heparin was included in the hybridization solution (HS); hybridization and early washes were at 65°C, and the secondary antibody was preadsorbed to Drosophila embryos and diluted 1 :2000. Mounted preparations were examined under Nomarski optics with a Zeiss Axiophot microscope. The fro message, which hybridizes to and is detected by the antisense RNA probe, was highly expressed in pre-meiotic primary spermatocytes during their growth period but was not detectable at earlier stages in germline stem cells or mitotically dividing spermatogonia. The transcript persisted in meiotic cells but was not detectable in differentiating spermatids. The transcript was present in both fzo¹ /fzo¹ and fzo²/Df(3R)P20 testes. A control sense-strand sense fro RNA probe did not hybridize in cells of any stage.

Example 7: Fusion Proteins Derived from Fzo and Antibodies Thereto

Production of a Fzo Fusion Protein. An expression construct that encodes and directs the production of a fusion protein containing a 6-histidine (6xHis) tag , an amino acid sequence that binds Nickel (Ni), and the carboxy terminal 115 amino acids of Fzo was prepared as follows. A 0.5 kb SamHI/H/ndlll restriction fragment from the fro cDNA1 (the Hinύ\\\ site is from the Bluescript SK- multiple cloning site) was subcloned into vector pQE30 (QIAGEN). Expression of the fusion protein was induced, protein was harvested under denaturing conditions, and the 6xHis/Fzo fusion protein was purified on a Ni-NTA column with imidazole elution as described in the QIAexpressionist manual (QIAGEN). Aliquotsfrom each column fraction were electrophoresed on an SDS-polyacrylamide gel and stained with Coomassie Blue. Fractions with the highest degree of purity and highest concentration of the fusion protein were dialyzed in 1.5 M urea, 0.1 M Na phosphate, 0.01 M Tris-HCI, and 500 mM NaCI, pH 7.

Generation of Antibodies to Fzo Amino Acid Sequences. After dialysis, precipitated fusion protein was emulsified in complete Freund's adjuvant and injected into rabbits; standard schedules were followed for booster injections and serum collections (Berkeley Antibody Company, Richmond CA). The fusion protein formed a precipitate at urea concentrations below 4 M. Approximately 500 mg and 250 mg of the fusion protein were used for initial and booster injections, respectively.

Western Blots. Protein extracts were electrophoresed in 10% SDS-polyacrylamide gels and transferred to nitrocellulose filters in methanol buffer using standard methods. Testis extracts were prepared by placing freshly dissected testes into a solution containing 8 M urea, 0.1 M sodium phosphate, and 0.01 M Tris-CI, pH 7 and mixing thoroughly until all proteins had dissolved. One volume of SDS-PAGE sample buffer was added, and the samples were then boiled for 15 minutes and spun at top speed in a microcentrifuge for ten minutes. Eight testes' worth of extract supernatant was loaded in each lane. Whole fly extracts were prepared by homogenizing flies in the above urea solution, with further treatment as above. One fifth of a fly's worth of extract was loaded per lane.

Filters were blocked in Blotto with 0.1 % Tween 20 and incubated for two hours at room temperature in either control (preimmune) or anti-Fzo^604"718 serum (fourth bleed) diluted 1 :1200, or mouse monoclonal anti-actin antibody (Amersham) diluted 1 :100, in the blocking solution. Subsequent washes, incubation with horseradish peroxidase-conjugated secondary antibodies, and detection were performed as described in the ECL Western Blotting manual (Amersham). The secondary antibodies were diluted 1 :7000 (anti-rabbit IgG) or 1 :2000 (anti- mouse IgG). Immunofluorescence. Fly testes were prepared for immunofluorescence staining as described in Hime (1996) J. Cell Sci. 709:2779-2788. The resulting slides were incubated overnight at 4°C in preimmune or anti-Fzo^604"718 serum (fourth bleed) diluted 1 :150 in PBTB (phosphate-buffered saline with 0.1 % Triton X-100 and 3% bovine serum albumin), washed four times at room temperature in PBTB, and incubated for one hour at 37°C in FITC- conjugated anti-rabbit IgG (Jackson Labs) diluted 1 :200 in PBTB (0.5 mg/ml RNAse was included for samples to be later stained with propidium iodide). Slides were washed 4x10 minutes in PBTB and mounted in 85% glycerol, 2.5% N-propyl gallate. For some samples, 1 μg/ml DAPI (Sigma) was included in the second PBTB wash after incubation in secondary antibody; for others, 1 μg/ml propidium iodide was included in the mounting medium. Samples were examined using epifluorescence on a Zeiss Axiophot microscope; images were collected with a Photometries cooled CCD camera (courtesy of Bruce Baker). Emissions from different fluorochromes on the same sample were collected separately and overlaid using Adobe Photoshop.

Antibodies raised against a fusion protein containing the C-terminal 115 residues of Fzo detected a wide doublet in Western blots of wild type testis extracts. The Fzo protein was partially resistant to extraction and denaturation and under standard denaturing conditions ran predominantly in streaky globs at an apparent molecular weight of 67 kD. When subjected to additional boiling and denaturants, some of the protein shifted up to a smoother band at an apparent molecular weight of 82 kD, the predicted size for Fzo. The wide Fzo doublet was greatly reduced in fzo¹ and fzo²/Df(3R)P20 mutant testes and, as expected, increased in testes from wild type flies carrying six copies of a feo⁺ transgene.

Very low levels of Fzo or a cross-reacting protein with the same mobility pattern were apparent in extracts from whole adult agametic males, indicating expression in somatic tissues. Similar low levels of this protein were seen in agametic females, ovaries, and embryos of all stages and were unaffected in fzo²/Df(3R)P20 whole male extracts. However, in fzo¹ /fzo¹ males, the Fzo protein is highly expressed in somatic tissues despite its downregulation in fzo /fzo¹ testes. The fzo¹ mutation therefore appears to affect tissue specific expression of Fzo. Extracts from whole wild type males carrying six copies of the fro⁺ transgene showed more Fzo protein than agametic males but considerably less than fzo¹ /fzo¹ males. Fzo may normally be expressed weakly in somatic tissues; it is therefore possible that Fzo is required elsewhere in addition to the testis and that the fzo¹ and fro² alleles cause loss of function in the male germline only. Alternatively, the weakly-expressed somatic protein could be a Fzo isoform or an unrelated cross-reacting protein.

The Fzo protein is associated with mitochondria in early spermatids during a short time period that spans mitochondrial fusion. The anti-Fzo^604"718 antiserum stained onion stage Nebenkerns brightly in wild type spermatids but was undetectable or present at greatly reduced levels in fzo¹ /fzo¹ or fzo²/Df(3R)P20 testes, respectively. A wild type fro transgene in a fzo /fzo¹ mutant background restored detectable Fzo protein to the Nebenkern.

The Fzo protein was associated with mitochondria in wild type spermatids during a narrow developmental window corresponding to the time that Fzo function is required. Mitochondria align on the spindle equator throughout meiotic divisions; however, the Fzo protein was not detected associated with mitochondria until the end of meiosis, during telophase II. In post-meiotic haploid spermatids, Fzo was associated with aggregating mitochondria and was detected at highest levels associated with onion stage Nebenkerns. The Fzo protein was detected at lower levels associated with eariy elongation-stage mitochondrial derivatives and was not detected associated with more elongated mitochondria. In testes from flies carrying multiple different copies of the fro⁺transgene, the anti-Fzo antibody stained mitochondria much more brightly but with a similar time course to wild type. Preimmune serum did not stain wild type testes.

Example 8: Fzo Homologs and Mitofusins The fro transcription unit encodes a predicted protein of 718 amino acids (see SEQ ID NO:2 and GenBank accession number U95821 ). The predicted Fzo protein sequence was compared using the BLAST program to nucleotide sequences in GenBank and dbEST translated in all reading frames. Significant homology was detected to predicted proteins in mammals (Homo sapiens, Mus musculus and Rattus norvegica), worms (Caenorhabditis elegans), and yeast (Saccharomyces cerevisiae). Predicted proteins encoded by Fzo homologs were themselves subjected to homology searches as above. Alignments were done with the help of the CLUSTALW program using DNAstar software.

ORFs from C. elegans (U29244, ORF 14) and S. cerevisiae (Z36048) encode complete predicted proteins with 28% and 19% identity to Fzo, respectively. The S. cerevisiae predicted protein is itself 24% identical to a partial predicted protein from the thermophilic bacterium Caldocellum saccharolyticum (L18965 ORF 6), which is 1 1 % identical to Fzo. As the latter cells lack mitochondria, the C. saccharolyticum protein is unlikely to play the role of a mitofusin per se in its natural activties, although it may play a role in the fusion of bacterial cells and may function as a mitofusin if contacted with mitochondria. The C. saccharolyticum protein is, in any event, a Fzo homolog as defined herein. Genetic and biochemical analysis in S. cerevisiae confirms that the yeast homolog, feo7p, is required for mitochondrial fusion (Hermann et al. (1998), J. Cell Biol. 743:359-374; Rapaport et al., (1998) J. Biol. Chem. 273:20150-20155). Depletion of Fzol p activity from yeast cells rapidly blocks mitochondrial fusion and causes defects in mitochondrial distribution. In time, loss of Fzopl activity leads to cells depleted in mitochondrial DNA (as determined by DAPI staining) and a petite growth phenotype. Moreover, the yeast Fzo homolog is associated with mitochondria as an integral membrane protein. As with the Drosophila gene, mutations in the predicted GTPase domains of the yeast fzo homolog destroys the ability of the protein to mediate mitochondrial fusion. The yeast Fzopl protein is thus both a Fzo homolog and a mitofusin.

Partial or complete reading frames for mouse, rat and human Fzo homolog proteins were initially assembled in silico from ESTs. These sequences are used to prepare probes and PCR primers that are used to isolate full-length cDNAs according to methods known in the art. The mammalian species have two forms of related Fzo homologs (which are designated as mitofusins herein due to their localization to mitochondria) which are referred to herein as Mfn1 and Mfn2. Furthermore, Mfn 1 -specific sequences are present on two RNA species of different sizes, suggesting that two related but distinct forms of the Mfn1 protein are expressed in some instances. These isoforms of Mfn1 are referred to as Mfnl a and Mfnl b.

Example 9: Expression Constructs Encoding Human Fzo Homologs and Mitofusins The BLAST search with the predicted Fzo protein identified a human cDNA clone from which a human fetal brain EST (GenBank T06373) is derived. The human cDNA clone was obtained from the American Type Culture Collection (ATCC, Manassas, VA) and its nucleotide sequence was determined according to standard methods.

The predicted amino acid sequences of hMfnl and hMfn2 are provided in the sequence listing. A plasmid comprising a hMfn2 cDNA sequence was obtained as a gift from the Kazusa DNA Research Institute (Chiba, Japan), and a cDNA encoding hMfn2 was prepared therefrom as follows. The plasmid was used as a template for polymerase chain reaction (PCR) in order to generate a hMfn2-encoding cDNA. Two synthetic oligonucleotides, having the following sequences, were used as primers in the PCR reactions:

5'-GACTCTAGAATGTCCCTGCTCTTCTCTCG (SEQ ID NO:27), and

5'- GCCCACTATCTGCTGGGCTGCAGG (SEQ ID NO:28). In the above descriptions of the PCR primers, sequences corresponding to hMfn2 sequences are underlined and the start (ATG) and stop (CTA, i.e., the reverse complement of TAG) codons for hMfn2 are doubly underlined. SEQ ID NO:27 includes an Xba\ restriction site (5'- TCTAGA) near its 5' end. The Taq-amplified PCR products were directly ligated ligated into an expression vector, pCR®2.1 (Invitrogen, Carlsbad, CA) that has ends complementary to Taq-amplified PCR products. The resultant plasmid, pCR-hMFN2 is both an expression construct (when transcription is driven by the T7 promoter, RNA molecules that encode hMFN2 are produced) and an antisense construct (when transcription is driven by the lacZ promoter, which is oriented in the opposite orientation from that of the T7 promoter, the resulting RNA molecules have the reverse complement of a sequence that encodes hMFN2).

Transient overexpression of hMfn2, or of an hMfn2-GFP fusion protein that has a N- terminal deletion that removes GTPase motif G1 (motifs G2, G3 and G4 are retained in this fusion protein) in mamallian cells caused severe aberrations in mitochondrial moφhology and distribution leading to colapse into a few mitochondrial aggregates per cell if expression was allowed to continue unchecked. This ability to promote mitochondrial fusion is one activity or function of mitofusins that is used to evaluate the properties of derivatives of Mfn proteins and to screen for or assay agents thought to enhance, modulate or inhibit mitofusin functions or activities. Expression of the hMfn2-GFP fusion protein in these cells did not affect the morphology of other cellular compartments, i.e., the Golgi network, microtubule network and endoplasmic reticulum, these other compartments having been visualized with antibodies directed to markers specific therefor (beta-COP, alpha-tubulin and PDI, respectively).

Example 10:

Functional Derivatives of Human Fzo Homologs and Mitofusins Preparation ofhMfn2 Fusion Proteins. Plasmid pCR-hMfn2 was digested with XΛol and EcoRI and electrophoresed on an agarose gel, and the restriction fragment, which had the predicted the predicted size of 2.3 kb, was purified from the gel using standard methods. Plasmid pEGFP-N1 (Clontech, Palo Alto, CA) was digested with Xho\ and EcoRI and mixed with the purified hMfn2-encoding restriction fragment. T4 DNA ligase was added to the mixture, which was incubated at 4°C overnight. Competent E. coli cells were transformed with the ligation mix, and several transformants were independently isolated. The expression construct DNA contained in transformants was purifed as in the preceding Examples and restriction mapped.

An isolate comprising an hMfn2-GFP expression construct was identified by virtue of having the predicted restriction map, and this construct was used in further studies. In the hMfn2-GFP fusion protein, the most caΦoxy terminal amino acid residue of hMfn2 (R757) is joined to a short linker sequence (ILQSTVPRARDPPVAT, SEQ ID NO:29), which is in turn directly joined to the most amino terminal amino acid (M1 ) of the GFP portion of the protein. When transiently expressed in murine 10T1/2 fibroblasts, the hMfn2-GFP fusion protein localized to mitochondrial compartments, as demonstrated by fluorescence microscopy. A hMfn2-myc fusion protein was prepared in similar fashion. In this derivative, the hMfn2 amino acid residues at the amino terminal end of the fusion protein are linked to a short polypeptide sequence that corresponds to an epitope derived from the human c-myc protein. Antibodies specific for this polypeptide epitope are available from a number of commercial sources and may be used in the detection (or affinity purification) of myc fusion proteins. During transient expression in COS-7 cells, the hMfn2-myc fusion protein co- localized with a fluorescent fusion protein (ATP-translocase-GFP) known to localize to mitochondria.

These mitofusin fusion proteins, and others like them, are tested for other biochemical and biological activities and functions (other than mitochondrial localization) of mitofusins using methods and compositions described in the present disclosure, as well as those described in Bereiter-Hahn et al., 1994, Microscopy Research and Technique 27:198-219; Kuroiwa et a/., 1994, Microscopy Research and Technique 27:220-232; Hales et al., 1997, Cell 90:121-129; Hermann et al., 1998, J. Cell Biol. 743:359-373; and Rapaport et a/., 1998, J. Biol. Chem. 273:20150-20155. Mitochondrial fusion and megamitochondria formation can be monitored according to a variety of methods known in the art using commercially available reagents such as, for example, MitoTracker™ CMXRos (Molecular Probes, Eugene, OR; Karbowski et al., 1999, Biochim. Biophys. Ada 1449:25-40).

A variety of functional derivatives of mitofusins, having different activities and functions of the mitofusins from which they are derived, are prepared, identified and characterized in this manner. Functional derivatives of one type are used as guides to prepare related functional derivatives. For example, deletion analysis of the mitofusin fusion proteins (see following section) is used to identify portions of mitofusin proteins that are necessary for mitofusin activties and functions such as, by way of non-limiting example, mitochondrial localization; ability to complement fro and fro7p mutations in D. melanogaster and S. cerevisiae, respectively; and the ability to promote mitochondrial fusion in vivo or in vitro, etc. Synthetic oligopeptides, having amino acid sequences corresponding to the omited segments of mitofusin deletion derivatives that lack a specific activity or function, are tested for their capacity to enhance, modulate or inhibit the activity or function. Oligopeptides (or other derivatives) having such properties are functional derivatives of mitofusins.

Deletion of Sequences Associated with the Transmembrane Domain. In order to see if carboxy terminal amino acid residues of hMfn2 are required for mitochondrial location of hMfn2, the following experiments were carried out. The hMfn2 reading frame includes an Λ/col site that straddles a codon (ATG) that encodes M603 of hMfn2. In the GFP expression vector pEGFP (Clontech), the portion of the expression vector that encodes the GFP polypeptide has an Λ/col site that straddles the start codon forthe GFP reading frame. When combined, these two Λ/col sites are in-frame with respect to each other. An expression construct comprising the nucleotide sequences for hMfn2-GFP was digested with Λ/col and mixed with Λ/col- digested pEGFP DNA. The mixture was treated with DNA Ligase, and used to transform cells. Several transformants were independently selected, and expression construct DNA contained therein was purifed. The DNAs were restriction mapped, and an isolate having the predicted restriction map was identified, designated phMfn2-GFPΔΛ/col(a), and used for further studies.

The reading frame in phMfn2-GFPΔΛ/col(a) results in the deletion of the amino terminus (amino acids 604-757) of hMfn2 and the fusion of M603 of hMfn2 to V2 of GFP. The putative transmembrane domain of hMfn2 includes amino acids 618-637 and is thus entirely deleted in the hMfn2-GFPΔΛ/col(a) gene product. The hMfn2-GFPΔΛ/col(a) gene product was transiently expressed in mouse 7077/2 fibroblasts, and the distribution of the fluorescent fusion protein in the cells was examined by fluorescence microscopy. Unlike the intact hMfn2-GFP fusion protein, which localizes to mitochondria, the hMfn2-GFPΔΛ/col(a) fusion protein was distributed throughout the cytosol. Thus, all or some of amino acid residues 604-757 of hMfn2 are required for its proper insertion into the mitochondrial membrane.

The hMfn2-GFPΔΛ/col(a) gene product is one example of a functional derivative of a mitofusin protein that is a hydrophilic and water soluble, rather than membrane-bound, protein. The hMfn2-GFPΔΛ/col(a) gene product retains the predicted GTPase domain that mutational analysis demonstrates is required for mitofusin function (Example 5), and is thus expected to retain the biochemical functions associated with the predicted GTPase domain. Because hydrophilic and water-soluble proteins are easier to incorporate into screening assays, including high throughput screening assays, in which reactions and binding events necessary for such assays take place in the aqueous phase, such water-soluble functional derivatives of mitofusins have distinct advantages over hydrophobic, membrane-bound forms of mitofusins.

Another functional derivative of hMfn2 is prepared as follows. The above-described plasmid encoding the hMfn2-GFP fusion protein is partially digested with Λ/col that has been diluted 10x or 100x in 1x reaction buffer for varying periods of time. The reactions are stopped by addition of phenol and or heat inactivation of Λ/col (65°C, 10 min.) and the partically digested DNAs are ethanol precipiated according to standard methods. The resuspended DNAs, as well as molecular weight markers and a control Λ/col digest that has been allowed to go to completion, are electrophoresed on agarose gels that are stained with ethidium bromide. Partial digests are identified, and the corresponding DNAs are diluted and treated with DNA ligase. The ligation mixtures are used to transform cells, and several colonies are independently isolated. The plasmid DNAs present in the isolated transformants are prepared and restriction mapped, and a deletion derivative that has lost the short DNA fragment present between the two Λ/col sites in the hMfnl reading frame is identified. The gene product encoded by this deletion derivative is designated hMfn2- GFPΔΛ/col(b). In the hMfn2-GFPΔΛ/col(b) gene product, amino acids 600-612 (having the sequence MVTGLASLTSRTS, SEQ ID NO:14) are deleted from the full-length hMfn2 sequence. The deleted amino acids are immediately adjacent to, and on the amino terminal side of, the transmembrane domain and may be necessary for localization of hMfn2 to the mitochondrial membrane. The localization of the hMfn2-GFPΔΛ/col(b) gene product is determined in the same manner as was that of the hMfn2-GFPΔΛ/col(a) gene product.

The above-described deletion derivatives of hMfn2, hMfn2-GFPΔΛ/col(a) and hMfn2- GFPΔΛ/col(b), are tested for their ability to function as dominant negative forms of hMfn2 as were the above-described missense mutants. The Λ/col deletion derivatives of hMfn2 are expected to retain the ability to bind other components of the molecular machinery that mediates mitochondrial fusion, but do not localize to the mitochondrial membrane. Accordingly, other mitochondrial fusion components that bind to the Λ/col deletion derivatives are directed away from their site of action within or at the surface of mitochondria and are thus unable to mediate mitochondrial fusion or do so with a limited efficiency.

Many mitochondrial proteins have an amino terminal sequence that is necessary for or stimulates mitochondrial localization. However, hMfn2(aa1-96)-GFP, a fusion protein in which amino acids 1 -96 of hMfn2 are fused to GFP (expressed from an expression construct prepared in similar fashion as that expressing hMfn2-GFPΔΛ/col), did not localize to mitochondria and was distributed throughout the cytosol. Further, a fusion protein in which amino acids 97-757 of hMfn2 are fused to a c-myc eptiope, hMfn2(aa97-757)-t77yc, was effectively localized to mitochondria.

Example 11 : Hybridization Analyses of Human Mitofusin (hMfn) Genes DNAs comprising sequences specific for hMfnl or hMfn2 were radiolabelled and used in Northern blot analyses of mRNA isolated from a variety of human tissues. The results indicate that there are two transcripts (approximately 4 and 6.5 kb in size as estimated by comparison to molecular weight standards) comprising hMfnl -specific sequences in a variety of tissues. Expression of hMfnl was most pronounced in heart and pancreas. In contrast, a single transcript approximately 5 kb in size was detected by hMfn2-specific probes, and its expression was most pronounced in heart and muscle. Labelled hMfnl and hMfn2 probes were also used to probe filters onto which human

RNAs from a variety of healthy and diseased cell types, and from tissues prepared at various times of development, in order to determine if the genes are differentially expressed in different diseases or at different stages of development. These experiments indicate that hMfnl is preferentially expressed in certain cancer cell lines and in fetal tissue.

Example 12: Informational and Structural Features of Fzo Homologs Fzo homologs, including mitofusins, comprise a variety of distinguishing informational features (e.g., nucleotide and amino acid sequences) and structural features (e.g., predicted coil or transmembrane regions). These features include the following.

GTPase Motifs. The Drosophila fzo gene encodes a conserved predicted GTPase required for mitochondrial fusion during spermatogenesis and associated with mitochondria only during a short time spanning the fusion event. Fzo contains four motifs common to virtually all known GTPases and conserved among Fzo homologs from mammals to yeast. Fzo is the first protein to be assigned a function in this new family of large predicted transmembrane GTPases. Within this family, the GTPase domain placement, motif spacing, and protein size (but not overall sequence) suggest a possible distant relationship with dynamins, which play a central biomechanical role in endocytic membrane trafficking.

Mutations predicted by analogy to diminish guanine nucleotide binding (see preceding Examples) do not affect localization of Fzo to mitochondria but eliminated or reduced its ability to mediate mitochondrial fusion. The fzo^K99T mutation, predicted to disallow key hydrogen bonds with the GTP β and γ phosphates, caused a severe loss of function phenotype. In contrast, the fro R249L mutation appeared to allow some mitochondrial fusion to occur, though not enough to restore normal sperm moφhology or motility. The Fzo arginine- 249 is part of the G4 motif and is predicted to contact the ribose moiety of GTP. Nearly all known GTPases have a lysine at positions analogous to Fzo R249 ; perhaps the conservative change to arginine in Fzo and its higher eukaryotic homologs reflects somewhat decreased importance of this residue for stable nucleotide binding. Mutations analogous to fro R249L in

R249L

H-ras reduce but do not eliminate GTP binding. It is possible that Fzo may have residual GTP affinity, explaining its partial function.

The region of highest homology between the Fzo, human fetal brain (50% identity to Fzo), C. elegans (49%), and S. cerevisiae (20%) predicted proteins is a 186 amino acid region containing four completely conserved signature motifs found in virtually all GTPases. The C. saccharolyticum predicted protein also contains these motifs. Outside the individual motifs there is no significant similarity to any known GTPase. However, the spacing between the GTPase motifs, their N-terminal placement in Fzo and homologs, and the overall predicted protein size are reminiscent of the dynamin family (Warnock and Schmid, 1996 BioEssays 18:885-893). The G2 motif (Bourne et al., 1991 Nature 349:117-127) consists only of a conserved threonine and has not been defined in dynamins. Both the Fzo and dynamin families have a conserved threonine exactly 20 residues beyond the G1 motif that may represent G2. Exemplary consensus amino acid sequences in the Fzo homologs are described in Table 2. Consensus Sequences for GTPase Domains in Fzo Homologs

Outside the GTPase motifs, Fzo and its homologs have moderate homology at the amino acid level (30%, 21%, and 19% identity beween Fzo and the human fetal brain, C. elegans, and S. cerevisiae homologs, respectively). However, the following results indicate that the Fzo homologs share several predicted structural features.

Transmembrane Domains and Flanking Regions. The sequences of Fzo and homologs were analyzed for predicted transmembrane domains and regions likely to form coiled coils with the TMpredict program and the COILS program, respectively. The Compute pl/Mw program on the ExPASy Molecular Biology Server was used to calculate predicted isoelectric points. All of the Fzo homologs have a predicted transmembrane domain nearthe C terminus embedded in a large (-35 amino acids) uncharged region interrupted by one to three basic residues.

Although transmembrane domains of proteins are sometimes difficult to identify based on homology alone, the sequences of hMfnl and hMfn2 are particularly homologous to each other. These sequences, IIVGGVIWKTIGWKLLSV (SEQ ID NO:45) in hMfnl and LWGGWWKAVGWRLIAL (SEQ ID NO:46) in hMfn2, are thus useful for identifying mitofusins from mammals, particulariy primates.

All the Fzo homologs have a predicted coiled coil region located to the C-terminal side of a predicted transmembrane domain, although the C terminal coiled coil probability in the S. cerevisiae homolog is lower than in other homologs. The C. elegans and S. cerevisiae homologs have strongly predicted coiled coil regions near their N termini. The analogous region in Fzo has a lower coiled coil probability (0.23) but is strikingly similar to the C. elegans homolog in its α helical projection, showing clear acidic, hydrophobic, and basic/polar faces. All four homologs are acidic overall between the N terminus and the transmembrane domain, with predicted isoelectric points (pi) near 5, and are basic in the C terminal tail, with predicted pis near 9.

Example 12 Control of mitochondrial morphology by human Mitofusin Material and Methods

Identification of human Mitofusin genes. Human Mfn1 sequences were identified by searching the human genome and EST databases. Appropriate primers were designed to amplify the ORF bearing part of the predicted hMfnl message by OneStep RT-PCR (Qiagen) from 50 ng of human heart poly-A⁺-RNA (Clontech, Palo Alto, CA). The resulting 2.2 kb cDNA fragment was recovered, subcloned into pCR2.1TOPO (Invitrogen, Carlsbad, CA) and sequenced. Human Mfn2 was identified as the KIAA0214 cDNA (Nagase ef a/.(1996) DNA Res. 3:321-329, 341-354) by searching the GenBank database with the Drosophila Fzo protein sequence using TBLASTN. The hMfn2/KIAA0214 cDNA clone (GenBank Accession No. D86987) is 4.55 kb in size and covers the entire 2274 bp ORF as well as additional sequences of the 5'- and 3'-UTR.

Generation of expression constructs. The human Mfn1 coding region was excised with

EcoRI from pCR2.1TOPO:: hMfnl and cloned into pEGFP-C2 (GFP- fusion constructs) or inserted into plRES-GFP (Clontech, Palo Alto, CA) to generate express untagged Mfn1. GFP-or myc-tagged forms of hMfn2 were generated by amplification of the hMfn2 ORF using appropriate primers and Turbo-Pfty-polymerase (Stratagene, La Jolla, CA). PCR products were subcloned into pcDNA3A-myc/His (Invitrogen, Carlsbad, California) or appropriate pEGFP expression vectors (Clontech, Palo Alto, CA) for N- or C-terminal GFP-fusion constructs. Deletion constructs were derived from these by dropping out fragments or using PCR to amplify defined regions of the hMfn2 cDNA. Missense mutations were introduced in hMfn2 using the Quickchange Mutagenesis kit (Stratagene, La Jolla, CA). The ANT-GFP expression construct (clone VLP32), and expression plasmids for HA-tagged human Dφ1 and mutant Drp1 ^K38A were used. Cell culture and transfection. COS-7 cells, HeLa cells, and mouse 10T1/2 and LA fibroblasts were grown in DMEM supplemented with 10% fetal bovine serum and 100U/ml streptomycin, 100 U/ml penicillin. Cells were seeded on LabTekll glass slides and transfected using SuperFect (Qiagen) per manufacturer's instructions. Pools of stably transfected cell populations expressing hMfn2-myc were prepared by incubating transfected COS-7 cells in DMEM +10% FBS and 0.8 mg/ml Geniticin, maintained in selection medium for several passages over a period of 10 weeks, then processed for immunochemistry.

Antibodies and indirect immunofluorescence. Primary antibodies were: mouse monoclonal anti- cytochrome C-oxidase I subunit 1 :60 (anti-COXI, Molecular Probes), rabbit polyclonal anti-mouse Tfam 1 :1000 (Larsson et al. (1998) Nat Genet. 18:231-236). mouse monoclonal anti-myc-9E10 1 :40, mouse monoclonal anti-HA 1 :500 (Covance, Richmond, CA), mouse monoclonal anti-protein disulfide isomerase 1 :50 (anti-PDI; Sigma), rabbit polyclonal anti-β-COP 1 :100 (Sigma) and mouse monoclonal anti-α-tubulin 4AI 1 :100. Mitotracker Red CXRos (Molecular Probes) was used to label mitochondria per manufacturer's protocol. About 19-24 hours after transfection, cells were rinsed twice with Dulbecco's PBS (phosphate buffered saline), fixed with 4% formaldehyde/PBS for 15 minutes at room temperature, then rinsed three times and permeabelized for 15 minutes in PBTD (PBS with 0.1 % Triton X-100 and 0.05% sodium deoxycholate), followed by 30 minutes block with PBTB (PBS with 0.1 % Trition X-100 and 3% bovine serum albumin). Slides were incubated with primary antibody two hours at 37°C, rinsed four times with PBTB, incubated two hours at 37°C in FITC-/TRITC/Rhodamine-conjugated anti-rabbit or anti-mouse IgG (Jackson Immunochemicals; 1 :200 in PBTB), washed extensively in PBS, and examined by epifluorecence on a Zeiss Axiophot microscope. Images recorded by CCD camera (Princeton Instruments, Trenton, New Jersey; IPLab Sofware, Spectrum Sofware Signal Analytics Corp.) and processed with Adobe (Adobe, Mountain View, CA) Photoshop.

Results

Two human genes encode Fzo homologs. Two human genes designated Mitofusin 1 and 2 (hMfnl and hMfn2) encode homologs of Drosophila fzo. The human Mfn1 and Mfn2 predicted proteins had highest homology to Drosophila fzo protein in the N-terminal, predicted GTPase-domain. Overall, hMfnl (741 amino acids) was 32% identical and 50% similar to Drosophila fzo protein, while hMfn2 (757 amino acids) was 33% identical and 54% similar to Drosophila fzo protein. The predicted hMfnl and hMfn2 proteins were 60% identical and 77% similar to each other, with the most extensive homology in the predicted GTPase domain and the least conserved regions in the N- and C-terminal ends. Expressed sequence tags (ESTs) homologous to the hMfnl and hMfn2 genes were also found in mouse, cow, Xenopus and zebrafish.

Mammalian Mitofusins associate with mitochondria. Human Mfn proteins co-localized with mitochondrial markers when expressed in stably transfected COS-7 cells (myc-tagged hMfn2) or after transient transfection with constructs encoding full-length hMfn2 protein fused to GFP in all cell types tested (HeLa, LA9, 10T1/2, COS-7). hMfn2 epitope tagged with myc and hMfnl untagged or tagged with N-terminal GFP (GFP-hMfn1 ) also associated with mitochondria when expressed after transient transfection.

Expression of human Mfn proteins alters mitochondrial morphology Mitochondrial morphology was dramatically altered in transiently transfected cells expressing hMfn2. The effect ranged from conversion of the normal dispersed distribution of punctate mitochondiato more reticular structures, to extensive perinuclear clustering. Perinuclear clustering of mitochondria was observed in over 90% of transiently transfected cells expressing either hMfn2-myc or hMfn2-GFP. Perinuclear clusters usually appeared to contain most of the mitochondria in the cell, based on counterstaining with an independent mitochondrial marker. The perinuclear clusters varied in size and shape from cell to cell, but were usually crescent- shaped aggregates or perinuclear rings of distinct fluorescent particles. Expression of C- terminai mychMfn2, N-terminal GFP hMfn2, and untagged or C-terminal GFP- hMfnl protein after transient transfection also resulted in perinuclear mitochondrial clustering. Mitochondrial morphology appeared unaffected in parallel control transient transfections with an ATP-translocase-GFP construct, which drives overexpression of a protein of the inner mitochondrial membrane. Although the shape of the clustered mitochondrial mass in hMfri2- GFP-tranfected cells was reminiscent of the Golgi apparatus, counterstaining with a Golgi marker revealed that the Golgi apparatus and the mitochondrial cluster in hMfn2-GFP expressing cells did not co-localize. Transient expression of hMfn2-GFP did not grossly affect the structure and subcellular localization of the Golgi apparatus, microtuble cytoskeleton, or endoplasmic reticulum, based on co-staining with anti-p-COP, anti-α-tubulin or anti-PDI. The hMfn2 GTPase is not required for mitochondrial clustering. Mitochondrial clustering after transient overexpression of hMfn2 may not be due to excessive mitochondrial fusion activity in transfected cells. Conserved residues in the predicted GTPase domain of Drosophila fzo protein and yeast Fzolp are required for mitochondrial fusion. A corresponding missense mutation in the predicted GTPase domain of hMfn2 (hMfn2^{K 09T}- GFP) did not reduce mitochondrial clustering or association of the fusion protein with mitochondria, suggesting that the predicted GTPase activity is not responsible for mitochondrial clustering. In addition, N-terminal deletion constructs encoding human Mfn2 or mouse Mfn1 lacking parts of the predicted GTPase domain (deletion of hMfn2 from N terminus through G1 ; deletion of mMfnl from N terminus through G3, still showed mitochondrial clustering upon transient expression.

Residues near the hMfn2 C-terminus, however, were required for mitochondrial clustering. Both the Mfn1 and Mfn2 predicted proteins have a high probability of forming a coiled coil structure in the C-terminal 20 to 30 amino acids based on the algorithms of Lupas and Wolf (Wolf et al. (1997) Protein Sci. 6:1179-1189), a feature conserved in Drosophila Fzo. Truncated hMfn2 lacking the last C-terminal 54 residues and fused to GFP (hMfn2^1_703- GFP co-localized with the mitochondrial marker COXI when expressed in COS-7, HeLa or 10T1/2 cells after transient transfection. However, mitochondrial clustering was strongly reduced in hMfn2^1_703-GFP transfected cells compared to cells transfected with full-length hMfn2-GFP, suggesting a role for the predicted coiled-coil domain in perinuclear mitochondria! clustering.

A G TPase dependent hMfn function in mitochondrial morphology revealed by altering activity of the predicted mitochondrial fission protein Drp1. In vegetatively growing yeast cells, mitochondrial fusion is balanced by ongoing mitochondrial fission mediated by action of the dynamin related protein Dnm1. To test if effects of overexpression of hMfn on mitochondrial fusion might be masked by rapid mitochondrial fission, we co-expressed hMfn fusion protein and a mutant form of the mammalian dynamin related protein Dφ1 predicted to dominantly interfere with function of endogenous Drp1 activity. Expression of hMfn2 caused formation of long strings and networks of mitochondria in the presence of the dominant mutant human Dφ1^K38A protein. Formation of the mitochondrial strings and networks was blocked by a point mutation in a critical GTPase motif of hMfn2, suggesting that they require Mfn GTPase activity. Transiently expressed, HA-tagged wild type human Dφ1 protein was distributed ubiquitously in the cytoplasm of COS-7 cells transiently transfected with HA-Dφ1 and the mitochondrial marker ATP-Translocase-GFP (ANT-GFP). Expression of wildtype HA-Dφ1 after transient transfection did not visibly affect mitochondrial morphology compared to transfection with the ANT-GFP mitochondrial marker alone. Co-expression of wild-type HA- Dφ1 with wild-type hMfn2-GFP resulted in mitochondrial clustering resembling that observed after expression of Mfn2-GFP alone. HA-Drp1 partially co-localized with the clustered mitochondria.

A point mutation in a key GTPase residue of Drp1 resulted in an altered subcellular distribution of the Dφ1 protein. HA-Dφ1^K38A appeared in dot-like structures after transient transfection, in contrast to the diffuse cytoplasmic distribution observed for the wild type fusion protein. In addition, transient transfection with the HA-Dφ1^K38A construct affected mitochondrial moφhology, resulting in formation of perinuclear aggregates.

When hMfn2-GFP was co-expressed with GTPase-mutant HA-Dφ1^K38A, mitochondria formed thin GFP-positive tubules extending from the perinuclear mitochondrial cluster toward the cell periphery. In addition, many cells displayed an interconnected network of tubular GFP-positive structures. The GFP-positive threads and networks were mitochondria based on co-staining with anti-COXI.

Formation of mitochondrial threads and networks depended on a wild type hMfn2- GTPase domain. Mutation of a key predicted GTPase residue in hMfn2 blocked formation of mitochondrial threads and networks upon co-transfection with Drp1^K38A, resulting in mitochondrial clusters indistinguishable from those in cells expressing hMfn2-GFP alone. The HA-tagged Drp1^k38A mutant fusion protein appeared in dot-like structures in the hMfn2^κ109T-GFP/HA-Dφ1 ^K38A doubly transfected cells but did not appear associated with the perinuclear mitochondria.

An internal region ofMfn2 is sufficient for mitochondrial targeting. Little is known about the pathways governing targeting of nuclear encoded proteins to the mitochondrial outer membrane. To identify sequences required for association of hMfn proteins with mitochondria, we tested a variety of hMfn2 fusion proteins expressed after transient transfection. Although analysis of the hMfn2 sequence by PSORT and MITOPROT indicated a predicted mitochondrial targeting signal at the N-terminus, the N-terminal 96 amino acids of hMfn2 was neither sufficient nor necessary for association with mitochondria. hMfn2^97"757- myc protein co-localized with mitochondrial markers and caused perinuclear clustering of mitochondria when expressed after transient transfection.

A 93 amino acid internal fragment of hMfn2 containing the predicted transmembrane domain (TM) and C-terminal flanking sequences was sufficient to target a GFP reporter to mitochondria, although a small amount of GFP fluorescence also appeared throughout the cell. Expression of hMfn2^{6 0"703}-GFP did not cause perinuclear mitochondrial clustering. The predicted transmembrane domain contains two stretches of hydrophobic residues separated by one (fzo) or two (Mfns) charged residues. This bipartite hydrophobic region is followed by a predicted alpha-helical region characterized by a conserved stretch of charged residues. When the bipartite hydrophobic region was deleted from otherwise fulMength hMfn2-GFP, the protein no longer localized to mitochondria, as expected if the region acted as a transmembrane domain. Although the hydrophobic region is relatively short, it appeared important, as changing three hydrophobic amino acids into charged residues abolished mitochondrial targeting of the full-length fusion protein (hMfn2^RRE-GFP). Instead, the GFP reporter was distributed diffusely in the cytoplasm. GFP fluorescence also appeared high in the nucleus forthe hMfn2^RRE-GFP and hMfn2^{Δ60IW5 3}-GFP constructs, either because the cells were thicker over the nuclear region or because the mutant hMfn2^RRE-GFP was preferentially localized to the nucleus. In contrast, substituting other hydrophobic residues for these same amino acids (construct hMfn2^AAL-GFP) did not affect mitochondrial targeting. Mutations in the AH-region that changed conserved charged residues to non-charged residues did not abolish either targeting of the hMfn2-GFP fusion protein to mitochondria or formation of perinuclear mitochondrial clusters. Although many precursor proteins are imported into mitochondria by an amino-terminal cleavable pre- or targeting sequence, internal targeting sequences have been described for other nuclear encoded mitochondrial outer membrane proteins, including mammalian OMP25, which also has a C-terminal transmembrane domain containing region capable of correct mitochondrial localization.

The functional similarities between the yeast and Drosophila Fzo proteins suggest that mammalian fro homologs also mediate mitochondrial fusion. Consistent with this hypothesis, we found that expression of the human Mfn2 protein by transient transfection in cells also expressing a dominant interfering form of Drp7 caused formation of long strings and extensive networks of mitochondria. Formation of the mitochondrial cables and networks was dependent on a wild type GTPase domain of hMfn2, suggesting a central role for GTPase activity in the mechanism of action of the Mfn proteins, as in their homologs yfeo7 and Drosophila fzo. The mitochondrial networks closely resembled those formed by mitochondrial fusion in dnml mutant yeast cells due to unopposed activity of the Mfn homolog yFzolp. Dnm1 encodes a yeast homolog of human Dφ1 required for mitochondrial fission.

In mammalian cells, as in yeast, the size and moφhological arrangement of mitochondria is due to a dynamic balance between Mfn dependent mitochondrial fusion and Dφ1 dependent mitochondrial fission. Overexpression of Drp1^K38A may interfere with mitochondrial fission by blocking function of the endogenous Dφ1 protein. Accordingly, the formation of mitochondrial cables and networks observed in the doubly transfected cells appears to have resulted from excessive mitochondrial fusion due to overexpression of Mfn protein combined with decreased mitochondrial fission due to interference by the mutant Drp1.

In cells transfected with wild type hMfn alone or co-transfected with hMfn2 and wild type Drp1 , continuous mitochondrial fission may mask the effects of overexpression of hMfn by rapidly cleaving up any long mitochondria that form. In this case, a second, GTPase independent role of hMfn proteins in mitochondrial moφhogenesis or function may be revealed in the perinuclear aggregation of mitochondria. The predicted C terminal coiled coil domain and other parts of the Mfn protein could be involved in mechanisms essential for normal mitochondrial distribution, such as transport or positioning of mitochondria or lipid metabolism of mitochondrial membranes. Disruption of the mouse kinesin motor protein KIF5B also results in perinuclear mitochondrial clustering. Likewise, transient overexpression of the mammalian mitochondrial outer membrane protein OMP25 leads to perinuclear aggregation of mitochondria, possibly due to recruitment of synaptojanin 2A to mitochondria and local modulation of inositol phospholipid metabolism. Alternatively, overexpression of hMfn proteins could cause perinuclear clustering if the high levels of the protein expressed after transient transfection block mitochondrial import channels, causing defects in import or function of other mitochondrial proteins.

Biochemical analysis of the yeast Fzolp homolog indicated that the protein is anchored to the mitochondrial outer membrane with both its N-terminal GTPase domain and C-terminal ends exposed to the cytoplasm. If hMfns are associated with mitochondria via similar topology, the bipartite hydrophobic region could pass through the mitochondrial outer membrane and back out. The carboxy terminal predicted coiled-coil region (trimer-type for hMfnl ; dimer-type for hMfn2) could form an integral part of the fusion mechanism, reminiscent of membrane fusion by influenza virus Hemagglutinin or the SNARE protein complex. If so, we propose that the Mfn GTPase domain displayed on the cytoplasmic face of the mitochondrial outer membrane could interact with the C-terminal coiled coil region to regulate mitochondrial fusion.

All publications and patent applications cited in this specification are herein incoφorated by reference as if each individual publication or patent application were specifically and individually indicated to be incoφorated by reference. The citation of any publication is for its disclosure prior to the filing date and should not be construed as an admission that the present invention is not entitled to antedate such publication by virtue of prior invention.

Although the foregoing invention has been described in some detail by way of illustration and example for purposes of clarity of understanding, it will be readily apparent to those of ordinary skill in the art in light of the teachings of this invention that certain changes and modifications may be made thereto without departing from the spirit or scope of the appended claims.

Claims

WHAT is CLAIMED is:

1. An isolated nucleic acid molecule comprising a sequence encoding a Mfn2 polypeptide.

2. The isolated nucleic acid molecule of Claim 1 wherein said Mfn2 comprises the amino acid sequence set forth in any one of SEQ ID NO:4; SEQ ID NO:10 or SEQ ID NO:16.

3. The isolated nucleic acid of Claim 2, wherein said nucleic acid comprises the sequence set forth in any one of SEQ ID NO:3; SEQ ID NO:9; or SEQ ID NO:15.

An isolated nucleic acid encoding a mammalian mitofusin polypeptide.

5. The isolated nucleic acid of Claim 4, wherein said mammalian mitofusin polypeptide comprises the amino acid sequence set forth in any one of SEQ ID NO:6; SEQ ID NO:10; SEQ ID NO:14 or SEQ ID NO:16.

6. The isolated nucleic acid of Claim 4, wherein said nucleic acid comprises a sequence set forth in any one of SEQ ID NO:5; SEQ ID NO:7; SEQ ID NO:9; SEQ ID NO:11 ; SEQ ID NO:13; SEQ ID NO:15; and SEQ ID NO:12.

7. An expression construct comprising a transcriptional initiation region functional in an expression host, a nucleic acid having a sequence of the isolated nucleic acid of Claim 1 or Claim 4 under the transcriptional regulation of said transcriptional initiation region, and a transcriptional termination region functional in said expression host.

8. A cell comprising the expression construct of Claim 6 integrated into the genome of said cell.

9. A cell comprising a nucleic acid according to Claim 1 or Claim 4 as part of an extrachromosomal element or integrated into the genome of a host cell as a result of introduction of said expression cassette into said host cell, and the cellular progeny of said host cell.

10. A method for producing a mitofusin protein, said method comprising: growing the cell of claim 8, whereby said mitofusin protein is expressed; and isolating said mitofusin protein essentially free of other proteins.

11. A purified polypeptide composition comprising at least 50 weight % of the protein present as a mitofusin Mfn2 protein.

12. A purified polypeptide composition comprising at least 50 weight % of the protein present as a mammalian mitofusin protein

13. A functional derivative of a mitofusin Mfn2 encoding nucleic acid.

14. A functional derivative of a mammalian mitofusin-encoding nucleic acid.

15. A functional derivative of a mitofusin Mfn2 protein.

16. A functional derivative of a mammalian mitofusin protein.

17. The functional derivative of a mitofusin protein according to either of claims 15 or 16, wherein said functional derivative is a mitofusin fusion protein, a soluble form of a mitofusin protein or a soluble form of a mitofusin fusion protein.

18. A nucleic acid encoding the functional derivative of a mitofusin protein of claim 17.

19. An antibody specific for a mammalian mitofusin or a Mfn2 polypeptide.

20. A method of screening for biologically active agents that modulate mitofusin function, the method comprising combining a candidate biologically active agent with any one of:

(a) a mammalian mitofusin polypeptide, a Mfn2 polypeptide, or a functional derivative thereof; (b) a cell comprising a nucleic acid encoding a mammalian mitofusin polypeptide, a Mfn2 polypeptide or a functional derivative of said nucleic acid; or

(c) a non-human transgenic animal model for mitofusin gene function comprising a mammalian mitofusin polypeptide, a Mfn2 polypeptide; and determining the effect of said agent on mitofusin function.

21. A method of enhancing membrane fusion in a population of membrane-bounded entities, the method comprising introducing a mammalian mitofusin polypeptide, a Mfn2 polypeptide, or a functional derivative thereof, into said membrane under conditions that permit fusion of said entities.

22. The method according to Claim 21 , wherein said membrane-bounded entities are mitochondria.

23. A method of identifying an agent that reduces the fertility, or increases the mortality rate, of an arthropod pest comprising combining a candidate biologically active agent with any one of:

(a) a mitofusin polypeptide, or a functional derivative thereof;

(b) a cell comprising a nucleic acid encoding a mitofusin polypeptide or a functional derivative of said nucleic acid; or

(c) a non-human transgenic animal model for mitofusin gene function; and determining the effect of said agent on mitofusin function.

24. The method of claim 23, wherein said arthropod pest is an insect.

25. An agent identified by the method of claim 23.

26. A method of identifying an antifungal agent comprising combining a candidate biologically active agent with any one of: (a) a mitofusin polypeptide, or a functional derivative thereof;

27. An agent identified by the method of claim 26.

28. A method of identifying an agent that enhances, modulates or inhibits mitochondrial fusion, comprising combining a candidate biologically active agent with any one of:

(a) a mammalian mitofusin polypeptide, or a functional derivative thereof;